EPA-600/2-76-098a
April 1976
Environmental Protection Technology Series
BURNER DESIGN CRITERIA FOR CONTROL OF
NOx FROM NATURAL GAS COMBUSTION
Volume I
Data Analysis and Summary of Conclusions
3)
\
industrial Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into five series These five broad
categories were established to facilitate further development and application of
environmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The five series are:
1. Environmental Health Effects "Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research performed to develop and
demonstrate instrumentation, equipment, and methodology to repair or prevent
environmental degradation from point and non-point sources of pollution. This
work provides the new or improved technology required for the control and
treatment of pollution sources to meet environmental quality standards.
EPA RE VIEW NOTICE
This report has been reviewed by the U. S. Environmental
Protection Agency, and approved for publication. Approval
does not signify that the contents necessarily reflect the
views and policy of the Agency, nor does mention of trade
names or commercial products constitute endorsement or
recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield. Virginia 22161.
-------
EPA-600/2-76-098a
April 1976
BURNER DESIGN CRITERIA
FOR CONTROL OF NOX
FROM NATURAL GAS COMBUSTION
VOLUME I. DATA ANALYSIS AND SUMMARY OF CONCLUSIONS
by
D.R. Shoffstall
Applied Combustion Research
Institute of Gas Technology
ITT Center--3424 South State Street
Chicago, Illinois 60616
Contract No. 68-02-1360
ROAPNo. 21BCC-029
Program Element No. 1AB;014
EPA Project Officer: David G. Lachapelle
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, NC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
ABSTRACT
Volume I of this final report gives a detailed presentation and analysis
of trials conducted with natural gas to determine the relationship between
combustion aerodynamics and pollution emission characteristics of industrial
burners. Three types of burners were studied (kiln, ported baffle, and
movable-vahe boiler) based on their relative gas load and estimated total
industrial emissions. Experimental measurements carried out on a pilot-
scale furnace included a baseline characterization of each burner and
variation of primary operating parameters (air preheat, air/fuel ratio,
firing rate, heat-release rate, position of gas nozzle in burner block, and
air swirl intensity). Additional emissions data were gathered for suspected
control conditions (fuel injector design, flue gas recirculation, fuel/air
momentum ratio and burner block angle). This volume also contains a
detailed description of the experimental facility and sampling probes used
to collect the data.
A companion publication, Volume II,gives a complete discussion of the
procedure used to select the test burners. Included also are detailed flame
characterizations of base-line operations assembled from in-the-flame
temperature, gas species, and flow direction data analysis. Similar in-the-
flame studies were made for control conditions which minimized emissions
for each burner type. All raw data collected from the input-output trials
are also included.
111
-------
TABLE OF CONTENTS
ABSTRACT iii
LIST OF FIGURES vii
LIST OF TABLES . xvii
ACKNOWLEDGMENT xix
INTRODUCTION 1
BURNER SELECTION RATIONALE 3
DESCRIPTION OF EXPERIMENTAL BURNERS 6
KILN BURNER 6
SWIRL BAFFLE BURNER 10
UTILITY BOILER BURNER 10
DESCRIPTION OF FURNACE TEST FACILITY 14
EXPERIMENTAL PLAN 29
GENERAL FLAME CHARACTERISTICS 31
KILN BURNER 34
BAFFLE BURNER 70
UTILITY BOILER BURNER 110
APPENDIX. DATA CORRELATION 167
-------
LIST OF FIGURES
Figure No. Page
1 Cutaway view of experimental kiln burner 7
2 Divergent gas nozzle 8
3 Combination gas nozzle 9
4 Assembly drawing of baffle burner 11
5 Boiler burner 12
6 Boiler burner air register vanes 13
6a Rectangular test furnace 15
7 Removable sidewall furnace panels; for interior 16
flame probing
8 Overall system schematic diagram of rectangular 17
test furnace system
9 Radiant tube preheater for main furnace combustion 18
air
10 Flue-gas cooler 20
11 Control room facility and analytical instrumentation 22
12 Gas-sampling probe head for nonparticulate flue gases 25
13 Modified IFRF temperature probe 26
14 General probe holder 28
15 Flame types tested 32
16 Flame geometry and luminosity of kiln burner 36
17 Flame geometry and luminosity typical of kiln burner 37
18 In-the-flame profiles of kiln burner using combination 38
nozzle with 30$ axial and 70% radial injection; 3.2^
primary air
19 Normalized NO concentration as a function of percent O2 39
in the flue (excess air) for the combination nozzle kiln
burner using 30^ axial and 70$ radial injection and 3.2$
primary air
VII
-------
LIST OF FIGURES, Cont.
Figure No. Pag<
20 In-the-flame profiles for combination nozzle kiln 43
burner with 30$ axial and 70$ radial injection, with
13$ EFGR
21 Normalized NO concentration as a function of O2 in 44
the flue (excess air) for combination nozzle kiln burner
with 30$ axial and 70$ radial injection using EFGR (13$)
22 Normalized NO concentration as a function of O2 in the 46
flue (excess air) for the combination nozzle kiln burner
with 30% axial and 70% radial injection with a 1130°C
wall temperature
23 Normalized NO concentration as a function of O2 in the 47
flue (excess air) for combination nozzle kiln burner
with 30% axial and 70% radial injection using 13 % EFGR
with reduced wall temperature, 1150°C
24 Normalized NO concentration as a function of O2 in the 48
flue for combination nozzle kiln burner with 30% axial-
70% radial injection with 6% primary air and a 1310°C
wall temperature
25 Normalized NO concentration as a function of O2 in the 49
flue for combination nozzle kiln burner with 30% axial
and 70% radial injection with 6% primary air and a
1150°C wall temperature
26 Comparison of NO formation for 6% and 3% primary 51
air on the combination nozzle kiln burner using 30%
axial and 70% radial injection
27 Normalized NO concentration as a function of O2 in the 52
flue (excess air) for combination nozzle kiln burner
with the nozzle in the exit position using 3. 5% primary
air and 30% axial-70% radial injection
28 Normalized NO concentration as a function of O2 in the 54
flue (excess air) for combination nozzle kiln burner
using 14% axial and 86% radial injection, 3. 5% primary
air and 1345°C walls
29 Normalized NO concentration as a function of O2 in the 55
flue (excess air) for combination nozzle kiln burner
using 6.6% primary air; 14% axial-86% radial injection
and 1345°C walls
Vlll
-------
LIST OF FIGURES, Cont.
Figure No. Pag
30 Normalized NO concentration as a function of O2 in 56
the flue (excess air) for combination nozzle kiln burner
using 3.5$ primary air; 14$ axial--86$ radial injection
and 1320°C walls; with nozzle in exit position
31 Normalized NO concentration as a function of O2 in 57
the flue (excess air) for combination nozzle kiln burner
using 3.2$ primary air and 0.0$ axial gas injection
32 Normalized NO concentration as a function of O2 in 58
the flue (excess air) for combination nozzle kiln burner
in the exit position and using 0.0$ axial gas injection
33 Normalized NO concentration as a function of O2 in ' 60
the flue (excess air) for combination nozzle kiln burner
with reduced gas input (1900 CFH) and 30$ axial gas
injection
34 Normalized NO concentration as a function of O2 in 61
the flue (excess air) for combination nozzle kiln burner
fired with 1800 SCFH of gas and 0.0$ axial injection
35 Normalized NO concentration as a function of O2 in 63
the flue (excess air) for divergent nozzle kiln burner
using 3.5$ primary air; 2700 SCFH gas input and 1320°C
walls
36 Normalized NO concentration as a function of O2 in 65
the flue (excess air) for divergent nozzle kiln burner
using 3.5$ primary air; 2700 SCFH gas and (cooled)
1145°C walls
37 Normalized NO concentration as a function of O2 in 66
the flue (excess air) for divergent nozzle kiln burner
operated with (cooled) 1150°C walls and 9.5$ primary
air
38 Normalized NO concentration as a function of O2 for 71
the IFLB burner with a standard gas nozzle at gas
inputs of 3070 and 2005 SCFH
39 IFLB burner with standard fuel nozzle 72
40 IFLB burner with divergent nozzle 72
IX
-------
LIST OF FIGURES, Cont.
Figure No. Page
41 Nozzle'positions tested for the baffle burner 73
42 In-the-flame profiles along the axis of a baffle burner 75
for typical operating conditions
43 Normalized NO concentration as a function of flue O2 for 76
the IFL/B burner with a standard gas nozzle and 15*^ and
30$ FGR
44 Normalized NO concentration as a function of flue O2 for 78
the IFLB burner with a standard gas nozzle at a wall
temperature of 965°C
45 Normalized NO concentration as a function of flue O2 for 79
the IFLB burner with a combination gas nozzle and radial
injection
46 Normalized NO concentration as a function of flue O2 for 81
the IFLB burner with a combination gas nozzle and axial
and radial injection
47 Normalized NO concentration as a function of flue O2 for 82
the IFLB burner with the various gas nozzles in different
positions
48 In-the-flame profiles for the axial fired baffle burner 84
in the controlled operating condition
49 Normalized NO concentration as a function of flue O2 for 86
the IFLB burner with standard, divergent, and axial gas
nozzles
50 Normalized NO concentration as a function of flue O2 for 88
the SFLB burner with a standard gas nozzle
51 Normalized NO concentration as a function of flue O2 for 89
the SFLB burner with a standard gas nozzle at wall
temperatures of 1450° and 1050°C
52 Normalized NO concentration as a function of flue Ob for 91
the SFLB burner with a standard gas nozzle and I5y> and
25$ FGR
53 Normalized NO concentration as a function of flue O2 for 92
the SFLB burner -with the various gas nozzles
54 Normalized NO concentration as a function of flue O2 for 94
the SFLB burner with standard, divergent, and axial gas
nozzles
-------
LIST OF FIGURES, Cont.
Figure No. Page
55 Normalized NO concentration as a function of wall 97
temperature for the SFLB burner with a standard gas
nozzle
56 Normalized NO concentration as a function of flue O2 for 98
the IFL/B burner with a low-velocity gas nozzle
57 Normalized NO concentration as a function of flue O2 for 100
the IFLB burner with low-velocity, high-velocity and
divergent gas nozzles
58 Normalized NO concentration as a function of flue O2 for 101
the IFLB burner with standard, divergent, and axial gas
nozzles
59 Normalized NO concentration as a function of flue O2 103
with 25% primary air
60 Normalized NO concentration as a function of flue O2 for 105
the LNO-I burner under luminous-flame operating
conditions
61 Normalized NO concentration as a function of flue O2 for 106
the LNO-I burner with 15$ primary air
62 Ring fuel injector 112
63 Nozzle heads for fuel gun injector 114
64 Nozzle positions tested for the movable-vane boiler 115
burner
65 Secondary combustion air deflector plate 117
6 5a 60-degree gun nozzle in exit position 118
65b 30-degree ring nozzle flame in deflector position 118
66 In-the-flame profiles of boiler burner operated under 119
typical conditions
67 Normalized NO concentration as a function of flue 120
O2 for the movable-vane boiler burner with a
60-degree gun nozzle
XI
-------
LIST OF FIGURES, Cont.
Figure No. Page
68 Normalized NO concentration as a function of flue O2 122
for the movable-vane boiler burner with a 60-degree
gun nozzle, 1. 5% and 2. 5% FGR
69 Normalized NO concentration as a function of flue O2 123
for the movable-vane boiler burner with a 60-degree
gun nozzle in different positions
70 Normalized NO concentration as a function of flue O2 125
for the movable-vane boiler burner with a composite
plot of gas nozzles
71 Normalized NO concentration as a function of flue O2 127
for the movable-vane boiler burner with a 60-degree
gun nozzle in different positions and a 15-degree vane
angle
72 Normalized NO concentration as a'function of flue O2 128
for the movable-vane boiler burner with a 60-degree
gun nozzle in different positions and a 45-degree vane
angle
73 Normalized NO concentration as a function of flue O2 129
for the movable-vane boiler burner with a composite
plot of gas nozzles at a 45-degree vane angle
74 Normalized NO concentration as a function of flue O2 130
for the movable-vane boiler burner with a 60-degree
gun nozzle in different positions and a 60-degree vane
angle
75 Normalized NO concentration as a function of flue O2 131
for the movable-vane boiler burner with a composite
plot of gas nozzles at a 60-degree vane angle
76 Tangential/radial velocity ratio as a function of vane 132
angle
77 Normalized NO concentration as a function of tangential/ 134
radial velocity ratio for the movable-vane boiler burner
•with a 60-degree gun nozzle
78 In-the-flame profiles of boiler burner using 60-degree 136
gun nozzle
79 Normalized NO concentration as a function of wall 138
temperature for the movable-vane boiler burner with
a 60-degree gun nozzle
xii
-------
LIST OF FIGURES, Cont.
Figure No.
80 Normalized NO concentration as a function of secondary
air temperature for the movable-vane boiler burner
with a 60-degree gun nozzle
81 Normalized NO concentration as a function of flue O2 141
for the movable-vane boiler burner with a 30-degree
ring nozzle in different positions
82 Normalized NO concentration as a function of flue O2 142
for the movable-vane boiler burner with a 30-degree
ring nozzle in different positions and a 15-degree vane
angle
83 Normalized NO concentration as a function of flue O2 143
for the movable-vane boiler burner with a 30-degree
ring nozzle in different positions and a 45-degree vane
angle
84 Normalized NO concentration as a function of flue O2 144
for the movable-vane boiler burner with a 30-degree
ring nozzle in different positions and a 60-degree vane
angle
85 Normalized NO concentration as a function of tangential/ 146
radial velocity ratio for the movable-vane boiler burner
with a composite plot of nozzles
86 Normalized NO concentration as a function of wall 148
temperature for the movable-vane boiler burner •with a
30-degree ring nozzle
87 Normalized NO concentration as a function of secondary 149
air temperature for the movable-vane boiler burner with
a 30-degree ring nozzle
88 Normalized NO concentration as a function of flue gas 1 51
recirculation percentage for the movable-vane boiler
burner with a 30-degree ring nozzle
89 Normalized NO concentration as £L function of flue O2 1 52
for the movable-vane boiler burner with a 30-degree
gun nozzle
90 Normalized NO concentration as a function of flue O2 154
for the movable-vane boiler burner with a composite
nozzle plot and a 15-degree burner block angle
Xlll
-------
LIST OF FIGURES, Cont.
Figure No. Pag*
91 Normalized NO concentration as a function of flue O2 155
for the movable-vane boiler burner with a composite
nozzle plot, a 15-degree burner block angle and a
1 5-degree vane angle
92 Normalized NO concentration as a function of flue Oz 156
for the movable-vane boiler burner •with a composite
nozzle plot, a 15-degree burner block angle and a
45-degree vane angle
93 Normalized NO concentration as a function of flue O2 157
for the movable-vane boiler burner with a composite
nozzle plot, a 15-degree burner block angle and a
60-degree vane angle
94 Normalized NO concentration as a function of tangential/ 158
radial velocity ratio for the movable-vane boiler burner
with a composite nozzle plot
95 Normalized NO concentration as a function of flue O2 159
and vane angle for the movable-vane boiler burner with
a 30-degree ring nozzle
96 Normalized NO concentration as a function of burner 161
block angle for the movable-vane boiler burner with a
30-degree ring nozzle
97 Normalized NO concentration as a function of tangential/ 162
radial velocity ratio and burner block angle for the
movable-vane boiler burner with a 30-degree ring nozzle
98 Normalized NO concentration as a function of flue O2 163
for the movable-vane boiler burner with a composite
nozzle plot, 60-degree vane angle and 45-degree burner
block
99 Normalized NO concentration as a function of adiabatic 173
flame temperature for the kiln burner
100 Normalized NO concentration as a function of 174
[1000/TADB( °K) ] for the kiln burner
101 Logarithmic ratio of NO/[O2] as a function of 175
[1000/TADB( °K) ] for the kiln burner
102 Normalized NO concentration as a function of 176
adiabatic flame temperature for the intermediate
flame length ported baffle burner
xiv
-------
LIST OF FIGURES, Cont.
Figure No. Page
103 Normalized NO concentration as a. function of 177
[1000/TA:QB( °K) ] for the intermediate flame length
ported bailie burner
104 Logarithmic ratio of NO/[O2] as a function of 178
[lOOO/T^^gl °K) ] for the intermediate flame length
ported baTfle burner
105 Normalized NO concentration as a function of adiabatic 179
flame temperature for the short flame ported baffle
burner
106 Normalized NO concentration as a function of 180
[1000/TArm( K) ] for the short flame ported baffle
-, -A.J-J.D
burner
107 Logarithmic ratio of NO/[O2] as a. function of 181
[lOOO/T.^-J °K) ] for the short flame ported baffle
-i J\i~) JD
burner
108 Normalized NO concentration as a function of adiabatic 183
flame temperature for the movable-vane boiler burner
109 Normalized NO concentration as a function of 184
[1000/TADB( °K) ] for the movable-vane boiler burner
110 Logarithmic ratio of NO/[O2] as a function of 185
[!OOO/TAD ( °K) ] for the movable-vane boiler burner
xv
-------
LIST OF TABLES
Table No. Page
1 Basic Burner Types and National NOX Emissions 5
Levels
2 Linear Equations for NOX Formation as a Function 40
of Oxygen Concentration, X
3 Synopsis of Data Collected for the Kiln Burner 68
4 Normalized NO as a Function of Nozzle Type, Block 85
Angle, and Excess O2 for the IFLB Burner With an
Air Preheat Temperature of 460°C
5 Normalized NO Concentration at 1 $ Excess O2 With 95
an Air Preheat Temperature of 460°C as a Function
of Baffle Type, Gas Nozzle Type, and Burner-Block
Angle
6 Synopsis of Data Collected for the Baffle Burner 108
7 Listing of Burner Operating Conditions as a Function 133
of Vane Angle for the 60-Degree Gun Nozzle (Gas
Input, 2996 SCFH; Exit and Deflector Positions; 1340°C
Wall Temperature; 30-Degree Burner-Block Angle;
460°C Secondary Air Preheat Temperature)
8 Listing of Burner Operating Conditions as a Function 145
of Vane Angle for the 30-Degree Ring Nozzle With a
30-Degree Burner Block (Gas Input, 2907 SCFH; Exit
and Deflector Nozzle Positions; 1369°C Wall Tem-
perature; 460°C Secondary Air Preheat; 2^ Excess
Oxygen)
9 Synopsis of Data Collected for the Boiler Burner 164
10 Burner Operating Conditions Used in Data Correlation 171
11 Conversion Table, English to Metric Units 186
xv 11
-------
ACKNOWLEDGMENT
The author would like to acknowledge the help of Mike Kerna, Lloyd McHie,
David Orchowski and Mike Peer who all contributed to the successful comple-
tion of this research. Thanks also goes to David W. Pershing and
Dr. J. O. L. Wendt, of the University of Arizona, and David Lachapelle, the
EPA Project Officer, for their many helpful discussions and contributions
made in the data analyses. The assistance extended by Cheryl Zrna and
Dennis Larson in preparation of this report is also appreciated.
xix
-------
INTRODUCTION
There is a current and vital need for significant reductions in the
pollutant emissions from stationary fuel combxistion sources. The Second
Annual Report of the Council on Environmental Quality (in 1971) stated that
stationary fuel combustion, which includes electric power generation, is the
second largest combustion-related NO emission source, contributing nearly
50% of the total national NO emissions. Only transportation ranked higher,
contributing 51. 4% of the nation's total NO emissions.
.X
The use of fuels for stationary combustion can be broken down into
household, commercial, and industrial,including the generation of electricity.
Within this breakdown, the industrial sector uses over 70^ of the annual
consumption of fossil fuel. Consequently, industry must be considered the
major source of pollutant emissions. Any development effort for reducing
these emissions will most certainly have a significant national impact.
Judging from the predictions of increased energy demand by industry, the
need for pollution control methods in industry will be even more critical
within the next 10 years.
There are two basic methods for reducing industrial combustion-
related emissions. The first method for decreasing pollution is to decrease
the amount of fuel consumed per unit of product or service. In this way,
emissions can be reduced in direct proportion to the decrease in fuel usage.
Both industry and energy suppliers are already actively engaged in devel-
oping process technology which will improve fuel utilization. There is a
substantial inducement to develop the means for better fuel utilization
because of both the pollution factor and reduced costs. However, considering
the rapid growth in the demand for products and services, the national con-
sumption of fossil fuels will still continue to grow and with it the total
emissions of pollutants. Better fuel utilization only helps to retard an
increase in total emissions.
The second method for decreasing atmospheric pollution from the
industrial combustion of fuels is to reduce the pollutant emissions per unit
of fuel burned. This method of solving the problem is completely compatible
with the first approach of reducing the amount of fuel burned per unit of
-------
product produced. This solution also is being actively pursued by a) switching
fuels (for example, by substituting low-sulfur oils or natural gas for high-
sulfur oil or coal), b) by developing and applying stack-gas cleanup processes,
and c) by modifying combustion systems. Of these three methods to reduce
pollution emissions, combustion modification is the least developed techno-
logically and is not yet widely available for commercial use by industry.
Only the utility power boiler segment of industry has implemented combustion
changes to significantly reduce emissions. Yet, new burner designs and
combustion systems specifically designed for low NO emissions offer the
most promising long-range solution.
The overall objective of this program is to develop technology so that
optimum low-emission (NO , CO, and HC) combustion systems can be
designed and widely used by industry. The work will establish the relative
controls available through the various burner designs and classical modi-
fication techniques for major burner classes. The specific result of this
study will be alternative control strategies for gas-fired systems.
To ensure that the results of this program have the most immediate
and greatest impact on the industrial emissions problem, three burner
types were selected for experimental study which represented a weighted
combination of the largest total gas usage and emission rates.
-------
BURNER SELECTION RATIONALE
In order to select three distinctly different types of burners to be
investigated, a classification system was developed which would categorize
a burner type by its combustion and heat-release characteristics. The
relative gas loads for each of these burner types were determined from gas
consumption statistics by industrial process and by an assessment of the
dominant burner type used for each process. The assessment of the pre-
dominant burner type by industrial process came from industrial burner
manufacturers. The gas load statistics by industrial process were available
in American Gas Association publications and from nonproprietary gas
supply and utilization studies conducted by IGT.
Obtaining relative NO emission rates by burner type was extremely
X.
difficult because of the lack of published data over the broad range of
industries covered. Therefore, each of the industry processes was assigned
into one of three categories of NO emission levels. It was assumed that
x #
the high emission processes emitted 0.5 Ib/million Btu; intermediate
emission processes, 0.25 Ib/million Btu; and low emission processes, only
0.05 Ib of NO /million Btu of fuel consumed. Obviously, very few processes
Ji.
and burner types produce exactly the quantity of NO of the group into -which
X
they were placed. However, for the purposes of this program evaluation,
this method provides a sufficiently good relative measure of the contribution
of each burner type to the national NO emission problem.
X.
Each industry process and burner type was placed in the appropriate
category based on available literature data or based on our expertise in the
area of NO emissions developed by the field testing of burners and the
X.
testing of scaled industrial burners in our laboratory. This latter method
was coupled with our knowledge of the firing rate, heat-release pattern,
percent excess air, and average temperature of the industrial process. We
therefore have data on the total gas load/yr and on the NO emission rate in
Ji
Ib/million Btu for each type of burner.
It is EPA policy to use Metric units; however, in this report English
units are occasionally used for convenience. See attached conversion
table.
-------
The selection of the three burners to be investigated was based on the
total pounds of emissions for each burner type on a national basis. Total
emissions for each burner type were determined by multiplying the gas
load/yr in Btu's by the NO emission rate in Ib/million Btu.
The results of this evaluation are presented as the identification of the
predominate burner type associated with each of the many fuel-consuming
industrial processes, a relative measure of the total national NO emissions
.X
by burner type, and the selection of three burners for study in this program
based on the highest national NO emission levels.
There are seven burner types identified by combustion characteristics.
These are shown in Table 1, along with the total national (estimated) NO
X.
emission levels established by this study. The three burners selected for
further experimental study were the 1) register burner, 2) the non-premix
gas-momemtum-controlled burner, and 3) the non-premix swirl burner.
These burners contribute significantly more NO to the national environment
than any of the others shown. The nozzle mix, nozzle premix, and fuel pre-
mix burners were grouped together because they are very often used
interchangeably by industry and therefore are difficult to evaluate separately
in terms of their NO emmisions.
x
The last category in Table 1, shown as "other," is made up of many
burner types, usually of a very specialized design or application. Any one
of these burners contributes very little NO .
' x
The three burners selected for further study may be more easily
recognized by their trade descriptions and applications. The register
burner is the typical design used on utility power boilers and large industrial
boilers. The non-premix gas-momemtum-controlled burner is more
commonly called a kiln burner and is used in open hearth steel furnaces,
glass melting, cement kilns, lime kilns, aluminum ore drying, and nonferrous
smelting furnaces. The non-premix swirl burner is sometimes called a
"baffle" burner or "large capacity" burner. It is the typical design used in
steel soaking pits, steel reheat furnaces, and other material heating processes
requiring temperatures up to about 2500°F.
A detailed classification by industrial burner type and their relative
NO emissions is presented in Volume II.
X.
-------
Table 1. BASIC BURNER TYPES AND
NATIONAL NO EMISSION LEVELS
x
NOX Emissions,
Burner Type 106 Ib/yr
1. Nozzle Mix
Nozzle Premix 84.3
Full Premix
2. Register Burner 831.2*
3. Flat Flame 7.6
4. Delayed Mixing 1.4
5. Non-Premix Gas Momentum-Controlled 441.0*
6. Non-Premix Swirl 92.4*
7. Other 11.9
Burners selected for further study in this research program.
-------
DESCRIPTION OF EXPERIMENTAL BURNERS
This program was conducted using three experimental burners simulating
a kiln burner, a swirl-type burner, and a utility boiler burner. Each of the
burners was designed for natural gas firing at a rate of 3.5 million Btu/hr.
The swirl-type baffle burner was full scale. The kiln and boiler burners
were aerodynamically scaled down by a factor of about 10. Each of the
burners was designed so that operating conditions could be readily changed
over a wider than normal range. Therefore, the burners can be used to study
"off-normal" conditions.
KILN BURNER
Figure 1 shows a cutaway view of the experimental kiln burner. The
combination of the burner block and housing simulate the firing end and air
duct entrance of a kiln aerodynamically scaled down from a 35 million Btu/hr
unit. The housing and block were velocity-scaled around a typical kiln
velocity of 10 ft/s and calculated for our input of 3500 SCFH of natural gas
and 10$ excess air.
The actual burner was designed with two different tip configurations. An
adjustable divergent gas orifice (Figure 2) can replace the radial flow tip
(Figure 3). Both tip designs have provisions for a purely axial flow or any
rate of axial-to-radial gas flow. An air flow at a velocity of 150-200 ft/s
around the gas nozzles is sized at about 20^ of the air put through the housing.
The gas nozzle tip opening is variable on both tip designs. It is designed
around a gas velocity of about 500 ft/s. However, the variable opening feature
allows variations in gas velocity from about 1000 ft/s (at maximum gas
pressure available at test site) to about 40 ft/s at the maximum-open setting.
The kiln burner (gas nozzle and secondary air annulus) is constructed so
it can be positioned flush with the inside wall of the furnace or pulled back
into the air housing up to 18 inches. Positioning the burner flush with the
inner furnace wall would be considered typical for most kiln designs.
-------
36 in.
t
\2 in
4
18.0 in.
•-^r
(EXIT)
8JPNER BLOCK WITH
15-deg DIVERGENCY
„ 2ljn
AIR FLOW CORRECTION
SCREEN (OPTIONAL)
(NORMAL) I
BURNER HOUSING
(SIMULATES IN SIZE '
KILN AIR PLENUM)
Jr
COMBUSTION AIR
INLET
AXIAL FLOW GAS INLET
RADIAL FLOW GAS INLET
VERNIER ORIFICE CONTROL
BURNER BLOCK CENTER LINE
PRIMARY AIR INLET
KILN BURNER SHOWN
IN WITHDRAWN AND
NORMAL POSITION
(SEE FIGURE 4, OF THIS REPORT)
A76040856
Figure 1. Cutaway view of experimental kiln burner
-------
Air Inlet
Gas Inlet-
Figure 2. Divergent gas nozzle
-------
sO
\XX\\\\\\\\\\N\\\\\\\\\\\\ \\ \\X\\\\
.LI
1.5"
Air Inlet
Axial Gas
•5
Rndinl
inlet
I
as T
L
Figure 3. Combination gas nozzle
-------
SWIRL BAFFLE BURNER
The swirl baffle burner used for experimental work was full scale. This
type of burner is found on many large process heating furnaces such as steel
reheating, batch glass melting, and tunnel kilns. It normally consists of a
centrally located gas nozzle surrounded by a baffle which has holes cast into
it for air flow (Figure 4). The burner is similar to the baffle burner used
by IGT in previous work for the EPA under Contract No. 68-02-0216.
The gas nozzle is designed for either radial flow, axial flow, or a
combination of both in any desired ratio. In addition, nozzles of various
diameters are available in order to vary axial gas velocity for a fixed gas
flow rate. The radial gas ports are designed for a velocity of about 600 ft/s
at a 3500 SCFH gas input. The axial nozzle is designed for a range of
velocities from about 230 to 1000 ft/s at a 3500 SCFH gas input, depending
on the diameter of the nozzle which is installed in the burner housing.
Varying the velocity of the radial or axial gas flow and/or the axial/radial
gas flow ratios will vary the flame characteristics.
The flame patterns of this burner can also be varied by changing the air
flow pattern and velocity. The air enters the burner block through six equally
spaced holes or ports in the ceramic air baffle. Baffles are available with
holes from 2.5 inches to 1.0 inches in diameter, which will vary air velocity
from about 50 ft/s to 330 ft/s. A swirl can also be imparted to the air flow
with baffles having the holes cast in at an angle. Three baffles with this
capability are available.
UTILITY BOILER BURNER (REGISTER TYPE)
Figure 5 shows a cross-sectional view of the experimental boiler burner
with a coal-firing attachment installed along with the conventional gas burner.
A vane register is used to impart swirl to the airstream as it enters the
combustion region. The vanes can be adjusted (Figure 6) from full open to
90 degrees closed. The full-open position causes fully radial air flow and
the longest flame.
10
-------
AIR PORTS
CENTRAL
GAS NOZZLE
"A" PIPE SIZE
GAS NOZZLE
SEAL BETWEEN -1 \,—
GAS NOZZLE 8 BAFFLE
AND BAFFLE 8 BODY WITH
R 8 I 3000 OR EQUAL
"D DIAM
A-34-415
NO SCALE
BAFFLE
a. LONG FLAME
b. SHORT FLAME
c. INTER. FLAME
AIR PRESSURE FOR
40,000 SCF/hr at 850 °F
3.25 in. we
18 in. we
14 in. we
"A"
1-1/4 in.
3/4 in.
1 in.
"B"
6 in.
5 in.
8 in.
"c"
2-3/8 in.
2-3/8 in.
3-7/8 in.
"D"
13 in.
16-1/2 in.
13 in.
Figure 4. Assembly drawing of baffle burner
11
-------
A-83-II99
Figure 5. Boiler burner
12
-------
,. AIR FLOW
A-34-414
Figure 6. Boiler burner air register vanes
The gas nozzle can be either of pure axial-flow or partial radial-flow
design. The radial design consists of an end-plugged pipe with six equally
spaced holes drilled around the circumference of the pipe and about 1/4 inch
from the end. The orifices or ports are sized for a gas velocity of about
300 to 600 ft/s at a 3500 SCFH gas flow. The burner block is only 6 inches
thick, which allows much of the combustion to occur within the actual furnace
enclosure.
13
-------
DESCRIPTION OF FURNACE TEST FACILITY
The experimental work was conducted on a rectangular furnace with a
25-sq-ft cross-sectional area and a 15-foot length. This furnace can be end-
or sidewall-fired at a rate of 4 million Btu/hr. The furnace is equipped for
in-the-flame sampling, preheated air, and flue-gas recirculation. (See
Figure 6a.) This furnace is capable of operating at temperatures up to
3000°F or as low as 1600°F at constant (maximum) gas input of 4 million
Btu/hr and up to 40$ excess air. Cooling is achieved with cooling coils
embedded in the refractory walls. The furnace is constructed completely
of 9-inch-thick cast refractories, with removable panels in one sidewall to
permit insertion of sampling probes (Figure 7). The overall furnace system
is shown schematically in Figure 8. The system is flexible enough that the
following operating parameters can be independently varied:
• Heat input, up to 4 million Btu/hr (8.0 million for certain burners)
• Air input, up to 40$ excess
• Heat losses to the furnace walls by changing flow in water-cooling
tubes cast into the refractories
• Combustion air temperature, up to 1000°F
• Flue gas recirculation capability, up to 35$ of combustion air
• Furnace pressure, up to +0.05 inch of water.
The combustion air for the main furnace can be preheated up to a
temperature of 1000°F with a separately fired radiant tube air preheater.
The radiant tube furnace (Figure 9) consists of an insulated airtight steel
chamber 4 feet high, 4 feet wide, and 16 feet long. As the combustion air to
be preheated passes through this chamber, it is heated by convection from
three 6-inch-diameter "hairpin" gas-fired radiant tubes.
The radiant tubes and refractory flow passages inside the preheater are
arranged to provide an S-shaped flow pattern, which maximizes residence
time for heating at the maximum allowable pressure drop (20 ounces) for
which the flow pattern will provide the necessary air flow of 75,000 SCFH.
14
-------
p-1;--
Figure 6a. Rectangular test furna,ce
-------
Figure 7. Removable sidewall furnace panels
for interior flame probing
-------
SAFETY SHUTOFF-
V33
5O-p«ig
GAS
PREHEATER COMBUSTION
AIR BLOWER-F5
MAIN GAS
SHUTOFF-VI
IOCK>«ifl
i V34 AIR
(PI4)
50-piig
/ GAS
MAIN GAS
SHUTOFF-
V34
BLEED VALVE-Vr7
METERING ORIFICE -08
V29
(TI8-T48)
WALL-COOLING TUBES,
METERING
PITOT-04
METERING ORtFICE-OI
FLUE DAMPER-
1 V28
(Til!
MANUAL SHUT-
FLOW CONTROL OFF-V7.V8
T2) VALVES-
SAFETY
SHUTOFF-V2
EXPERIMENTAL FURNACE
(P4) 02
METER ORIFICE
SECOND-STAGE PRESSURE
REGULATORS V5.V6
FIRST-STAGE PRESSURE
REGULATORS V3.V4
PUMP PI BYPASS
^
V2I
FLUE-GAS COOLER
AIR FLOW CONTROL-VII
i METERING
ORIFICE-07 IT8)
WALL-COOLING
SWITCHING VALVES
VI2.VI3
WALL-COOLING
BLOWER-F3
METERING ORIFICE-OS
(T5)
WALL-COOLING 1
PUMP-P2
WALL-COOLING WATER
HEAT EXCHANGER
ORIFICE CITY WATER
METER-06
RIVER-WATER
PUMP-PI
-, FLUE-GAS COOLER RIVER-WATER
- HEAT EXCHANGER FLOW CONTROL
\»LVES-VI4,VI5
MAIN COMBUSTION Ain
BLOWER-F4
FILTERED
AIR INTAKE
WALL-COOLING
PUMP-PS
DRAIN
Figure 8. Overall system schematic diagram
of rectangular test furnace system
-------
00
i
-4
i
-4
•
ft
AID
INI FT
(8 in.)
r
VIEWING
PANEL
FLOW BAFF
ft
AIR
INLET —
(8 in.)
— ^-
_E~
^-
^
IX*
f,
\
V^J
* " 0 * ° ° 6 ' * * 0
a o o
• ° o > ° ° " j ° ° ° ' e
. 0°,°'°'^ ' • •
^£d
tmim&mm^w&ikNn ~t\3tt. \to.'!> »-
~^
~ 16 ft
H
(
(12
REFRACTORY FLOW
\~PASSAGES
r
i RADIANT TUBE NO. 1
<• " ". r.'o"0» " % ^ • ' °Ul^U»0°
•jJ 0 0 - " . ' - . ' . <• ° 0 • • _*
§^ii;;;^wii;^^;i;ii RADIANT TUBE NO. 2 mmimm
* ^^
""
OT AIR
DUTLET
in. X 12 in.)
' 1
1
u^y V
*2JJ
t
-^
•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•
))
( | RADIANT TUBE NO. 3 —
^
— •>
(
b.
F
INTERNAL INSULATION
STEEL SHELL
RADIANT
FLUE
COLLECTOR
FLUE
COLLECTOR
D-34-463
Figure 9. Radiant tube preheater for main furnace combustion air
-------
The temperature of the air can be regulated by changing the heat input to
the radiant tubes. Ambient temperature air can be supplied by completely
shutting down the preheater or by directing the air through the preheater
bypass pipe. The bypass pipe was installed to allow working on the preheater
without shutting down the air supply to the main furnace. Air bypass is
achieved by selective switching of valves.
Flue products for recirculation back to the burner and main furnace are
obtained from the furnace itself. Flue products? can be withdrawn from the
furnace flue passage just prior to the main furnace flue damper. Up to
14,000 SCFH of flue products can be withdrawn from the flue, which provides
a 30$ recirculation factor when the furnace is fired at 3.5 million Btu/hr with
20^ excess air.
The main furnace flue products are actually pulled from the flue
(Figure 8) by the suction in the inlet side of the main furnace combustion
air fan (F4). The flue products enter the recirculation withdrawal and
treatment system, at about 2800°F, through a short length of internally
insulated steel duct. These hot gases are cooled to about 125°F in a packed-
bed water cooler (Figure 10). Cooled city water (about 70°F) is sprayed down
on a bed of refractory packings as the hot gases pass up through the packed
bed. This cooling system lowers the water content of the flue gas from about
0.008 to about 0.007 Ib/cu ft, which, is the dew point of the gases at about
125°F. (The lost water content can be readded later in the system if
experimental conditions require this treatment.) The cooled gases then pass
through a flow-control shutoff valve (V25 in Figure 8). This valve controls
the flue-gas flow rate, which regulates the percentage of recirculated products.
This valve is interlocked to an outlet temperature sensor on the gas cooler.
If the outlet temperature of the gases exceeds 150°F, which would damage the
combustion air fan, the control valve (V25) shuts down. This stops the flow
of flue gases. Beyond the flue-gas control valve, the flue gases are mixed
with the required amount of air for combustion. A control valve (V24) regu-
lates the amount of air pulled in by the fan. The total amount of air for com-
bustion and flue products is metered with an orifice plate (O7) at the outlet
of the fan. The flue products and air then pass into the air preheater or
preheater bypass pipe.
19
-------
COOLING WATER SYSTEM
60 gpm AT 150 psig
SPRAY MANIFOLD
AND HEADS"
COOLED
FLUE
GASES
WIRE MESH LIQUID
DEMISTTER
PERFORATED
PLATE"
2 in. WATER LEVEL
PACKED REFRACTORY
BED-80% VOIDS
WATER OVERFLOW
TO DRAIN
HOT WATER
COLLECTION TANK
•40 in.
~I6 in.
•36 in.
-20 in.
I
HOT(2800°F)
FLUE GASES
Figure 10. Flue-gas cooler
20
-------
The water used in the flue-gas cooler is clean city water, which is
continuously recycled. A water flow of 60 gprn is supplied at 150 psig, by
a turbine pump, to a series of spray heads in the gas cooler. The hot (200°F).
spent water flows out of the cooler into an atmospheric holding tank. This
tank is equipped with a constant-level overflow to the building drain. In this
way, any condensed water from the combustion is removed and disposed of
in the sewer. The water in the holding tank is periodically treated with
sodium hydroxide to prevent acid buildup in the water due to condensing flue-
gas components. One such component removed by the flue-gas cooler is NO2.
The hot water from the water holding tank is cycled through an American
Standard heat exchanger capable of removing 1.5 million Btu/hr of heat from
the water. Cooling in the heat exchanger is provided by a flow of river water
at a rate of about 150 gpm at 80 psig. The river water is supplied by a
river adjacent to the test facility through a service pump (Pi) maintained by
IGT.
Figure 11 is an-overall view of furnace controls and the analytical
instrumentation package. The eq\iipment used for concentration measure-
ments of chemical species during this program are listed below; these
analyzers included a;
Beckman 742 Polarographic Oxygen (O2)
Beckman Paramagnetic Oxygen (O2)
Beckman NDIR Methane (CH4)
Beckman NDIR Carbon Monixide (CO)
Beckman NDIR Carbon Dioxide (CO2)
Varian 1200 Flame lonization Chromatograph (Total HC and Cj to C9)
Beckman NDIR Nitric Oxide (NO)
Beckman UV-NO2
Hewlett-Packard Thermoconductivity Chromatography, Hydrogen (H),
Nitrogen (N2), Argon (A2), CO, CO2, Cj to C5, Oxygen (O2)
Beckman Chemiluminescent NO-NO2
Data Integration System
This instrumentation package allowed concentration measurements of
the following major components: a) measurement of hydrocarbon compounds
G! to C9; b) independent check of NO-NO2 chemiluminescent with NDIR-NO
and NDUV-NO2; c) independent check of paramagnetic O2, polarographic O2,
NDIR-CH4, NDIR-CO and NDIR-CO2 with the respective chromatographic
species concentration; and d) measurement of hydrogen (H2), argon (A2), and
nitrogen (N2).
21
-------
Figure 11. Control room facility and analytical instrumentation
22
-------
The following is a general description of the measurement system used
for this program.
1. NO and NO2 Instrumentation
The chemiluminescent NO and NDUV-NO2 system was mounted in a
X.
roll-around cabinet that could be placed out at the furnace. This was im-
portant in minimizing sampling distances, which can affect accuracy. The
chemiluminescent unit -was equipped with a carbon converter. Test work by
IGT and others has demonstrated that in a reducing environment the carbon
converter maintains a better conversion efficiency than converters made of
stainless steel, quartz or molybdenum.
The instrumentation was calibrated by using both a permeation tube
with a controlled known release of NO and certified prepared cylinders
X.
of NO and NO2 gases.
The sample gas was drawn from the furnace through a special alumina
probe by a Dia-Pump Model 08-800-73 all stainless steel and Teflon pump
delivering approximately 0.4 CFM. (This sample delivery rate was
dictated by the requirements of the measuring instruments.) The sample is
immediately passed through a stainless steel large-particle filter. Both
the pump and filter were kept above 100°C to prevent condensation of the
water vapor inherent to combustion products.
2. CO, CH4, and CO2 Measurements
Nondispersive infrared analyzsers were used for carbon monoxide,
methane, and carbon dioxide measurements. These analyzers do not affect
the sample gas and can be operated in series. They were calibrated by using
certified gases with known concentrations of the species being determined.
The infrared analyzers require a completely dry sample. Therefore, the
sample was first passed through a water trap and a 3 A molecular drying
sieve. A small in-line filter was placed immediately after the drying tube
to trap particles of sieve that may be carried over by the gas stream.
23
-------
3. Oxygen Measurements
A portion of the "conditioned" sample gas is diverted from the NDIR
units to a Beckman Model 600 paramagnetic analyzer. A second oxygen
analyzer, a Beckman Model 742 polarographic, was used as a cross-check
on the oxygen concentration. The Model 742 analyzer has an advantage over
the paramagnetic in time response.
4. Chromatographic Measurements
Asa detailed gas analysis was required, the sample is fed to a Hewlett-
Packard 7620-A thermal conductivity chromatograph which permitted con-
centration evaluations of hydrogen, nitrogen, argon, oxygen, carbon monoxide,
carbon dioxide and hydrocarbons Cj to C5. To achieve separation of these
species a helium carrier gas was used in conjunction with a Porapak Q
column. Three temperature program rates were also required ranging
from —100°C to 300°C. A sample loop volume of 100 ml was used to insure
linearity in the hydrogen response for concentrations up to 60$.
For total hydrocarbon analysis a Beckman hydrocarbon analyzer was
used. A detailed hydrocarbon analysis could be made using a Varian 1200
flame ionization chromatograph. All chromatographic readings were
electronically integrated and printed out as a function of resolution time.
In addition to flue gas analysis, a major task of this program was to
map profiles of temperature, chemical species and flow direction for each
burner type. Modified designs of the International Flame Research
Foundation were used to construct probes which enabled this type of data
collection.
Figure 12 shows the assembly drawing of the gas sampling probe used
both in the flame and flue. To minimize NO2 reduction in the probe, an
alumina tube was inserted for the first 18 inches and was joined to Teflon
tubing to carry the gas sample to the analyzers.
Temperature data was collected using a suction pyrometer whose design
is illustrated in Figure 13. A Pt-Pt Rh 10^ thermocouple was used. The
efficiency of the pyrometer was measured at 96$ with a 25 second response
time.
24
-------
CSJ
l-l/2-in. 304 SS TUBING
•l/2-in. 304 SS TUBING
-3/4-in. 304 SS
1-3/4-in. 304 SS TUBING
-------
WATER
GAS
EXTRACTION
-»—GAS
COOL 1X0
'.JACKEV
THERMOCOUPLE
HOT JUNCTION'
ALUMINA
SHEATH'
SILLIMANITE
SHIELDS
SUCTION TIP FOR MEASUREMENTS IN
NATURAL GAS AND OIL FLAMES
REFRACTORY
CEMENT PLUG'
SUCTION TIP FOR MEASUREMENTS IN
F'tJi-VF-illZED-COAL FLAMES
Figure 13. Modified IFRF temperature probe
A-15-37
-------
The gas sampling probe, suction pyrometer and direction flow probe
were all designed to be installed in the general probe holder shown in
Figure 14.
27
-------
-0.0625 in.
THICK •
oo
2.50in.x|| go
2.00 in. x 16 go
l.75in.x||go
Figure 14. General probe holder
-------
EXPERIMENTAL PLAN
The experimental measurements involved — a base-line characterization
of flue gas as a function of primary operating variables, a detailed in-the-
flame investigation*at selected base-line conditions, a characterization of
emissions as a function of various emission control techniques and/or com-
bination of techniques, and, finally, an in-the-flame investigation for those
control techniques found to be most effective for NO control.
For each of the burners studied, base-line operation of that burner was
characterized by conducting input/output measurements at the flue. This
characterization included the effects of changing the following operating
variables:
a. Three (3) air preheat temperatures (ambient, 400°, and 800°F)
b. Five (5) air/fuel ratios (excess air between stoichiometric and
251*, in 5w increments)
c. Two (2) firing rates (1.5 and 3.0 million Btu/hr)
d. Two (2) heat-release rates (as controlled by wall cooling with
air, with one-half of the water load and with the full water load)
e. Fuel injector design (radial, axial, and a combination of both,
or in the boiler-burner case, convergent radial, ring injector
and divergent radial, and gun injector)
Each operating variable was evaluated independently, while holding all
others constant. Thus, there were 300 different sets of conditions investi-
gated for base-line operation of each burner type. In addition to the quan-
titative flue measurements, flame characteristics were documented
photog raphically.
Upon completion of the base-line characterization with input-output tests
at the furnace flue, an intensive flame analysis was conducted for each of
the burners studied. These investigations were made at a preselected set
of operating conditions. These detailed studies included the following:
a. Directional flow analysis (Approximately 10 radial scans were
made in order to precisely define the configuration of the interval
recirculating regions.)
29
-------
b. Chemical species concentrations (Five radial profiles were con-
ducted with the gas samples being analyzed for N2, O2, H2, CO,
CO2, NO, NO2, and individual hydrocarbons).
c. Temperature profile (A suction pyrometer was used for five
radial scans to get time-averaged temperatures).
Once the base-line characterization was completed, this information was
combined with known and potential emission control methods. The specific
objective was to develop operating criteria for minimum emissions for each
burner type. The following parameters were investigated:
a. Two (2) flue gas recirculation quantities (added to combustion
air at 15$ and 30$ concentrations).
b. Two (2) combustion air swirl intensities (if applicable, studied
at two intensities differing from the base-line condition).
c. Fuel/air momentum ratio (vary the gas and air velocities by a
factor of 2 independently and also decrease the velocity of the
gas by a factor of 2).
d. Three (3) burner-block analyses (using three different burner-
block angles which differed from each other by 15 degrees).
e. Two (2) fuel injector positions (vary location of fuel injection to
plane which is even with the front wall of the furnace).
There were 270 different sets of low-emission operating parameters inves-
tigated. As in the base-line characterization investigation, flame
characteristics were documented photographically.
In order to develop an understanding of how the control techniques are
limiting the formation of pollutants, a detailed flame analysis was conducted
for one set of operating parameters which had minimized the pollutant con-
centration at the flue. The experimental investigation was carried out in a
manner similar to that previously discussed.
30
-------
GENERAL FLAME CHARACTERISTICS
A general aerodynamic characterization of a flame can be made by
determining the different types of flow patterns that exist within a com-
bustion chamber. A detailed flow analysis of a confined flame reveals that
the front section of a combustion chamber can be divided into four zones:
primary jet, primary recirculation, secondary jet, and secondary recir-
culation. The primary and secondary jets contain only particles with a
forward flow direction (away from the burner), whereas the recirculation
zones contain gas particles moving in the reverse flow direction (back
toward the burner). The size, shape, and particle density of the recircula-
tion zones are determined by the velocity, the ratio of gas to air, the spin
intensity of the secondary jet, the burner-block angle, and, for the secondary
recirculation zone, the size and shape of the combustion chamber. Figure 15
shows the three types of flow patterns that were observed during the pollution
control trials.
A type I flow pattern is observed when the secondary jet has no tangential
velocity component (no spin) and the primary jet velocity is much larger than
the secondary jet velocity. Depending on the initial jet exit velocity from the
fuel injector, the flame can be either attached to or detached from the injec-
tor. A type I flow pattern also can be generated with a secondary jet with a
tangential velocity component (spin) if the burner-block divergent angle is
equal to or less than the angle of the secondary jet relative to the burner
axis. As a result, the burner block restricts the expansion of the secondary
jet and inhibits the formation of a primary recirculation zone.
A type II flow pattern is generated when the secondary jet has a tangential
velocity component large enough to cause the particles to adhere to and pack
tightly against the burner block. This packing creates a low or negative
pressure region in the center of the burner block. The pressure differential
between the furnace and the central region of the burner block causes gas
molecules to be pulled into this region and back toward the burner, thus
creating the primary recirculation zone. When the velocity of the primary
jet is greater than the velocity of the recirculating gases in the primary re-
circulation zone, the primary jet penetrates this reverse flow region and a
recirculation lobe occurs on each side of the burner axis. This flow pattern
is labeled type II.
31
-------
TYPE I
NO SWIRL
MO PRIMARY
RECIRCULATION
TVPE JI
LOW SWIRL INTENSITY
PRIMARY JET VELOCITY >
SECONDARY JET VELOCITY
4
/\
\5ECONDARY I
\ REC.
x. V >- 1
TYPE HE
HIGH SWIRL INTENSITY
SECONDARY JET VELOCITY >
PRIMARY JET VELOCITY
Figure 15. Flames types tested
32
-------
If the central region of the burner block has a large enough negative
pressure differential relative to the furnace, such that the velocity of the
recirculating gases is greater than the velocity of the primary jet, pene-
tration of the primary recirculation zone is not possible and a type III flow
pattern results. A type III flow pattern also can be generated mechanically
by introducing the fuel with only a radial velocity relative to the axis of the
burner.
33
-------
KILN BURNER
An assembly drawing of the kiln burner which was tested was shown in
Figure 1. The air plenum, burner block combination were momentum scaled
to simulate the firing end and air duct entrance of a 35 million Btu/hr kiln.
For a natural gas firing rate of 3000 SCFH, standard trial input volume, a
10$ level of excess air would have a velocity of 13 ft/s when preheated to
450°C. Two types of fuel injectors were designed to fit the experimental
kiln burner. Figure 2 illustrates the divergent nozzle which has a 45-degree
cone, whose base faces the furnace, surrounded by a 45-degree angle diver-
gent orifice. The position of the cone is adjustable relative to the divergent
orifice permitting control of the gas injection velocity. The combination
nozzle is shown in Figure 3. This design includes an adjustable radial orifice
and a fixed diameter axial orifice. The fuel input can either be totally axial,
totally radial, or any desired ratio of axial to radial. The normal operating
gas velocity during the trials was 500 ft/s; however, the variable orifice
feature permitted variations in the gas velocity for a 3000 SCFH input from
40 ft/s to 1000 ft/s. Both the divergent and combination gas nozzles are
surrounded by an annular duct for primary air which is used to help shape
the flame. The injection velocity of the primary air is 6200 ft/s when its
volume is 5$ of the total combustion in volume.
The kiln burner was constructed so that the gas nozzle can be positioned
flush with the inside wall of the furnace or pulled back into the burner block
up to 13 inches. Positioning the nozzle withdrawn 13 inches into the burner
block will be referred to as the normal operating position since both jets,
primary and secondary, are momentum scaled during the initial stages of
mass exchange. Positioning the nozzle flush with the inner furnace wall will
be labeled the exit position and permits an expansion of the secondary jet
into the larger diameter combustion chamber.
Detailed flame surveys were conducted during the kiln burner trials.
These surveys included flow direction, temperature and gas analysis. The
details of this survey work along with complete in-the-flame data and a
comprehensive listing of the input-output data are presented in Volume II of
this report. However, it is beneficial in understanding the interpretation of
the input-output data to take a level look at some in-the-flame results.
34
-------
BASELINE CONDITIONS
The baseline characterization of the kiln burner was made using the
combination gas nozzle with 30$ of the total gas input being injected axially
(810 SCFH) and 70$ of the gas being injected in the radial direction
(1890 SCFH). The primary air was determined to be 3.2^ using the
relationship —
/
^ _. . . . primary air (SCFH) v , „„
* Primary Air = secondar^ aTr (SCFH) + primary air (SCFH) X 10°
The wall temperature was 1330° C. Photographs illustrating flame geometry
and luminosity with these and other operating conditions are shown in
Figures 16 and 17.
For the baseline operating conditions, the in-the-flame data revealed a
type I flow profile. The gas species and temperature data collected along
the burner axis are presented in Figure 18. The high methane concentration
(27.2$) and low flame temperature (825° C) at the burner block indicates a
very flow rate of mass exchange between the fuel and combustion air. The
NO concentration at the burner block, exit is 17 ppm or only 6.4$ of the flue
concentration. Thus the major formation of NO (93.6$) occurs within the
combustion chamber.
Figure 19 presents normalized nitrogen oxide (NO) versus excess
oxygen (O2) as a function of secondary air preheat for the baseline operating
conditions. Measured CO levels greater than 500 ppm are listed by concen-
tration next to their corresponding data point. These graphs show a linear
relationship between the normalized NO and the excess oxygen. Straight line
equations representing the dependence of normalized NO on excess oxygen for
these and Other operating conditions are presented in Table 2. This type of
relationship in the range of excess oxygen being tested is characteristic of
laminar diffusion flames. The rate of mixing between the fuel and combus-
tion air helps determine the slope of the line. This linear relationship will
break down at a stoichiometric fuel-air ratio, however extrapolating the test
data to zero excess oxygen gives a normalized NO flue concentration referred
to in this report as "stoichiometric. " Although stoichiometric operation of
industrial burners is not practical due to incomplete combustion and this cal-
culated "stoichiometric" concentration is not exact, it does provide a.basis for
comparing the relative effectiveness of changes in burner and furnace
35
INSTITUTE OF GAS TECHNOLOGY
-------
UJ
o-
Figure 16. Flame geometry and luminosity of kiln burner
-------
LO
-J
Figure 17. Flame geometry and luminosity typical of kiln burner
-------
OJ
oo
TEMP°C
-i JOO-i
1500-
1300-
1100-
90C-
700-
50U-*
250-
200-
IOO-
ICC
DISTANCE FROM BURNER - CM
150
Figure 18. In-the-flame profiles of kiln burner using combination nozzle with
30$ axial and 70$ radial injection; 3.2$ primary air
-------
450-
400-
350-
300-|
6
8:
200-j
i so-;
100-!
50-
COMBINATION NOZZLE, KILN BURNER
GAS INPUT 2700 SCFH; 810 SCFH AXIAL
1890 SCFH RADIAL
3.2% PRIMARY AIR
V.ALL TEMPERATURE I330°C (AIR COOLING)
SECONDARY AIR PREHEAT AS LABELED
460 C
D
a,
88
Q, IN FLUE, %
Figure 19. Normalized NO concentrations as a function of percent O2
in the flue (excess air) for the combination nozzle kiln burner
using 30$ axial and 70$ radial injection and 3.2$ primary air
39
-------
Table 2. LINEAR EQUATIONS FOR NOX FORMATION
AS A FUNCTION OF OXYGEN CONCENTRATION, X
Preheat
* Equation Temperature, °C
19 NO = 59X +111 460
NO = 25X +68 252
NO =11.5X + 57 22
21 NO =24.7X +72 463
NO = 7X +24 235
NO =4.4X +15 22
22 NO =30.5X +112 465
NO = 13X +55 254
NO =3X +43 22
23 NO = 9.75X +63 460
NO =3.5X +36 242
NO =3.OX +13 22
24 NO =51.7X +78 438
NO =17.7X +52 220
NO = 6.6X +48 22
25 NO =33.6X +118 466
NO =24.IX +53 272
NO = 11.4X +31 22
27 NO =69.OX +132 . 445
NO =23.4X +65 220
NO =8.2X +37 22
28 NO =70.3X +92 457
NO =32.8X +75 247
NO =9.2X +72 22
29 NO =63.7X +93 466
NO =19.4X +67 288
NO =9.7X +43 22
30 NO =57.7X +150 457
NO =27.8X +53 237
NO =8.4X +48 22
31 NO =44.2X +158 447
NO =26.6X +53 220
NO =16.2X +59 22
32 NO =70.3X +142 459
NO =21.8X +71 249
NO =13.OX +48 22
33 NO =17.8X +36 456
NO =6.9X +23 248
NO =7.2X +7 22
34 NO =23X + 53 460
NO =13.4X +43 242
NO =10X +32 22
40
-------
Table 2, Cont. LINEAR EQUATIONS FOR NOX FORMATION
AS A FUNCTION OF OXYGEN CONCENTRATION, X
Preheat
* Equation Temperature, °C
35 NO =31. OX +303 477
NO =17.8X +178 243
NO =14.5X +60 22
36 NO =23.6X +63 456
NO =18.7X +59 250
NO =8.7X +41 22
37 NO =17.OX +58 452
NO =10.OX +70 244
NO =6.4X +50 22
Refers to corresponding Figure No. in this report.
X = Percentage oxygen (O2) in the flue, $.
41
-------
operating conditions controlling NO emissions. As an example, the
"stoichiometric" concentration of normalized NO for baseline operating
conditions at 460°C secondary air preheat is a factor of two larger than the
concentration at ambient temperature. In addition, the normalized NO con-
centration increases at a rate five times greater for a 460°C secondary air
temperature than for ambient temperature air.
EXTERNAL FLUE GAS RECALCULATION (EFGR)
As a baseline control case, the effect of external flue gas recirculation
was tested. The percentage of EFGR is determined using the relationship —
$ EFGR = EFGR (SCFH)
secondary air (SCFH) + primary air (SCFH) + fuel (SCFH) uu
Figure 20 illustrates in-the-flame data collected along the burner axis for
baseline operating conditions with the addition of 13$ EFGR to the secondary
air. The temperature at the burner block exit is 1130°C or 305°C higher
than the baseline temperature. This, in addition to the reduction in methane
concentration from 28.6$ for baseline conditions to 17.5$ for 13$ EFGR
indicates that more of the total combustion was occurring within the burner
block. The total combustibles have been reduced by 32$ when the EFGR
control flame at the burner block exit on the burner axis is compared to the
baseline flame. The flame length is 106 cm or is 43$ shorter than the
baseline flame. The flow profile indicates that the flame is a type I. The
NO concentration is 7 pprfx or 5.8$ of the flue concentration. This percentage
of NO formation within the burner block is almost identical to the 6.4$
measured for the baseline operating conditions. Thus in the furnace the
EFGR flame has formed 113 ppm while the baseline flame produced 248 ppm.
This factor of two suppression in NO concentration occurs due to lower flame
temperatures produced by the EFGR dilution effect. The flue gas temperature
for the EFGR flame is 1415°C compared with 1560°C for the baseline flame.
Figure 21 presents the NO concentrations measured as a function of
excess oxygen and secondary preheat with the addition of 13$ EFGR to the
secondary air. During preheat both the flue gas and secondary air were
blended and reached the same final preheat. The EFGR effect on "stoichio-
metric" NO was a reduction of approximately 40 ppm independent of secondary
42
-------
V
A 0
u>
1500-
1300-
1100-
900-
700-
500-*
100 -
80 -
60 -
40 ~
20 -
TEMPPC
ITOO-i 120-^, 8-\ JO
7-
5-
I -
26-
22-
ICO
DISTANCE FROM BURNER - CM
ISO
Figure 20. In-the-flame profiles for combination nozzle kiln burner
with 30% axial and 70% radial injection^ with 13% EFGR
-------
450:
400-
350-j
COMBINATION NOZZLE, KlLN BURNER
GAS INPUT 2733 SCFH' 879 SCFH AXIAL
1854 SCFH RADIAL
3.2 °/b PRIMARY AIR
WALL TEMPERATURE I257°C (AIR COOLING)
13% FLUE GAS RECIRCULATION
SECONDARY AIR PREHEAT AS LABELED
300
E
Q.
2 5°1
s
S 200-i
I50H
100-
50-
J5I6
J5000
•V" 235°C
— O-
2 3
0 IN FLUE, °/o
Figure 21. Normalized NO concentration as a function of O2 in the flue
(excess air) for combination nozzle kiln burner with 30$ axial
and 70$ radial injection using EFGR (13$)
44
-------
air preheat (Figure 19 vs Figure 21). Although there was a variation in the
slope reduction as a function of preheat, the average reduction factor was
2.7 ± 0.4.
Comparing the baseline case (Figure 19) with the 13$ EFGR curve
(Figure 21) at 460°C shows a 52$ reduction in the flue concentration of NO
(at 3$ O2) as a result of 13$ EFGR.
WALL TEMPERATURE
The data for Figure 22 were collected with the kiln burner operating in
the baseline conditions with the exception of wall temperature. By water
cooling tubes mounted in the furnace sidewalls, the temperature was de-
creased from 1330° to 1130°C. The flue concentrations of NO were decreased
by approximately 30$ (from the baseline case) due to the lower wall tem-
perature at 3 $ excess oxygen.
Using the equations of Table 2 to compare NO emission levels from
baseline operation to emissions with water-cooled walls reveals a 43$
reduction using ambient temperature combustion air. While the intermediate
(254°C) and high (465°C) preheat levels of combustion air resulted in res-
pective reductions in NO emissions of 34$ and 30$.
Combining the NO reduction techniques of EFGR and water cooling of
the furnace sidewalls results in the emission curves plotted in Figure 23.
This combined control technique results in "stoichiometric" reduction of
43$ for the 460°C preheat temperature and a 77$ reduction for ambient air
temperature. Likewise, the magnitude of the slopes are reduced by approxi-
mately 80$. Translated directly into concentration reduction at the 3$ excess
oxygen level with a 460°C preheat, the combination control techniques resulted
in an emission level of 92 ppm compared to 288 ppm for baseline operating
conditions. This represents a 67% reduction in normalized flue NO
concentrations.
PRIMARY AIR
To test the effect of primary air on the flue NO concentrations the primary
air was roughly increased by a factor of 2, from 3.2$to 6$. The data for 6$
primary air are presented in Figure 24 for air cooling of the furnace sidewalls
and Figure 25 for water sidewall cooling.
45
-------
450-,
40CH
COMBINATION NOZZLE, KILN BURNER
GAS INPUT 2700 SCFH; 810 SCFH AXIAL
1890 SCFH RADIAL
3.5% PRIMARY AIR
WALL TEMPERATURE 1130°C (WATER COOLING)
SECONDARY AIR PREHEAT AS LABELED
350-
300-1
I
250-
I
200-
150-
100-
50-
465°C
22 C
1 • r—
2 3
02 IN FLUE, %
Figure 22. Normalized NO concentration as a function of O2 in the flue
(excess air) for the combination nozzle kiln burner with 30% axial and
70% radial injection with a 1130°C wall temperature
46
-------
450-,
400-
350-
COMBINATION NOZZLE, KILN BURNER
GAS INPUT 2700 SCFH- 810 SCFH AXIAL
1890 SCFH RADIAL
3.5 % PRIMARY AIR
WALL TEMPERATURE II50°C (WATER COOLING)
\3% FLUE GAS RECIRULATIOM
SECONDARY AIR PREHEAT AS LABELED
300-
I
^
250
-j
I
QC 200
150-
100-
50-
460°C
-o —
2 3
1 IN FLUE. %
Figure 23. Normalized NO concentration as a function of O2 in the flue
(excess air) for the combination nozzle kiln burner with 30% axial and
radial injection using 13$ EFGR with a reduced wall temperature, 1150°C
47
-------
450-,
400
350-
300-
250-
Jzoo
o
150-
100-
50-
COMBINATION NOZZLE, KILN BURNER
GAS INPUT 2706 SCFH; 876 SCFH AXIAL
1830 SCFH RADIAL
6.0% PRIMARY AIR
WALL TEMPERATURE 1310 c (AIR COOLING)
SECONDARY AIR PREHEAT AS LABELED
0 IN FLUE, %
Figure 24. Normalized NO concentration as a function of O2 in the flue
for combination nozzle kiln burner with 30% axial-70% radial injection
with 6% primary air and a 1310°C wall temperature
48
-------
450!
400-
350
£ 300
cs-
Q 250-
t^
-J
I
o zoo
150-
100-
50-
COMBINATION NOZZLE, KILN BURNER
GAS INPUT 2700 SCFH; 810 SCFH AXIAL
1890 SCFH RADIAL
6.2 % PRIMARY AIR
WALL TEMPERATURE I I50°C (WATER COOLING)
SECONDARY AIR PREHEAT AS LABELED
XD2000
Oz IN FLUE, %
Figure 25. Normalized NO concentration as a function of O2 in the flue
for combination nozzle kiln burner with 30% axial and 70% radial injection
with 6% primary air and a 1150°C wall temperature
49
-------
A comparison of these data and baseline data at 3 $ excess oxygen are
made in Figure 26. Wall temperature has little effect on NO emission levels
with 6$ primary air until the secondary combustion air temperature exceeds
200°C. The lower wall temperature then results in a gradual reduction in the
NO emission level which reaches 17% at a secondary air preheat temperature
of 470°C. Increasing the primary air percentage from 3.2$ (baseline opera-
tion) to 6$ while maintaining a 1330°C wall temperature results in a 9$
reduction in NO emissions at a 470°C secondary air preheat temperature
which gradually increases with decreasing secondary air preheat temperature
until it reaches a maximum of 18$ at ambient operation.
NOZZLE POSITION
Figure 27 presents NO concentrations measured under baseline condi-
tions, but with the gas nozzle in the exit position (refer to Figure 1). For
an excess oxygen level of 3$, the NO concentrations arising from baseline
operation are similar to those with the nozzle in the exit position at com-
bustion preheat temperatures below about 220°C. The deviation bet-ween the
measured NO concentrations then begin to increase as a function of com-
bustion air temperature until at about 450°C the nozzle in the exit position
produced 27$ more NO than the baseline conditions. This could result from
the combustion air scrubbing additional heat off the walls of the furnace. For
the lower secondary air preheats, this additional temperature increase is
compensated by entrainment of secondary recirculation products. For the
higher secondary air preheats, the additional temperature is sufficient to
aid in raising the average flame temperature and increasing NO concentrations.
Although this behavior is only exhibited by the kiln burner, the input-output
nozzle position data plus the in-the-flame data presented earlier indicate that
there is an optimum flame length or distance for fuel burn-out producing
minimum NO emission levels.
AXIAL/RADIAL RATIOS
To determine the dependence of NO formation on the direction of gas
injection and combustion intensity within the burner block, the ratio of axial
to total fuel injection volumes was decreased from the 30% used in the base-
line measurements to 14$ and 0$. Ratios above 30$ were not studied because
of excessive flame length. Due to the slow fuel-air mixing rate, the flame for
ratios above 30$ extended into the furnace flue.
50
-------
6 % PRIMARY
I330°C WALL
100 200 300 400
AIR PREHEAT °C
500
Figure 26. Comparison of NO formation for 6$ and 3% primary air
on the combination nozzle kiln burner using 30$ axial and 70% radial injection
51
-------
450
400
350
300-
O"
•250-
200
150-
100-
50-
COMBINATION NOZZLE, KILN BURNER
GAS INPUT 2773 SCFH; 873 SCFH AXIAL
isoo SCFH RADIAL
3.596 PRIMARY AIR
WALL TEMPERATURE I330°C
(AIR COO'.ING)
NOZZLE EXIT POSITION
SECONDARY AIR PREHEAT
AS LABELED
/445°C
6000 D
5500
I 2 3
Q, IN FLUE, %
Figure 27. Normalized NO concentration as a function of O2 in the flue
(excess air) for the combination nozzle kiln burner with the nozzle in the
jf jf jf
exit position using 3.5% primary air and 30°'' axial-70% radial injection
52
-------
Data collected for 14$ axial to total fuel input are illustrated in Figures 28
through 30. Figure 28 shows operating conditions similar to baseline conditions
with the exception of the 14$ fuel injection ratio. The larger radial volume gas
input resulted in higher levels of NO emissions. This is due to higher com-
bustion intensity within the burner block producing higher flame temperatures.
Comparing the baseline case (Figure 19) with the 14% axial gas input
(Figure 28), straight line equations reveal that an increase in slope as well as
"stoichiometric" NO occurs only for the intermediate preheat level. At 3 $
excess oxygen, this is reflected as a 22$ increase in NO emissions compared
with 9$ and 5$ increases at ambient and 46o°C secondary air temperatures,
respectively.
Increasing the primary air from 3.5$ to 6.6$ of the total air input resulted
in the NO emissions presented in Figure 29. The increased volume of primary
air brought the emission levels baick in line with those measured for baseline
conditions.
Figure 30 illustrates the NO concentrations measured as a function of
excess oxygen for the 14$ by volume gas axial input with the nozzle in the
exit position. The exit nozzle position provides for a reduction in the flue
concentration measured for 3$ excess oxygen and air preheat levels up to
about 380°C. The reduction at 250°C was approximately 25$ with a resulting
concentration of 132 ppm. Above 380°C, the exit nozzle position leads to
increased concentrations of NO with a level of 325 ppm, or a 7$ increase above
the baseline conditions at a 460°C secondary air preheat. The explanation for
the increase in NO emissions is due to the "scrubbing" heat transfer to the
combustion air postulated earlier in the nozzle position section.
Figure 31 shows the data for total radial gas injection. In general, the
NO concentrations measured were less than those with 14$ axial injection.
Compared with the baseline operating conditions, with no preheat, the total
radial gas injection results in 18$ higher emission at a level of 108 ppm;
however, for preheats above 125°C, the total radial injection produces slightly
lower (~5 ppm) levels of emissions. Data were collected for NO flue concen-
trations as a function of excess oxygen, with total radial injection and the
nozzle in the exit position. These results are illustrated in Figure 32.
53
-------
450
400
350
300
8:
8
Nj
-J
I
250
200-
150-
100-
50-
COMBINATION NOZZLE. KILN BURNER
GAS INPUT 2691 SCFH; 368 SCFH AXIAL
2323 SCFH RADIAL
3.596 PRIMARY AIR
WALL TEMPERATURE I345°C (AIR COOLING)
SECONDARY AIR PREHEAT AS LABELED
'457°C
Q, IN FLUE, %
Figure 28. Normalized NO concentration as a function of O2 in the flue
(excess air) for the combination nozzle kiln burner using 14$ axial and
86% radial injection, 3.5$ primary air and 1345°C walls
54
-------
450-
400-
350-
300-
250-
200-
150-
100-
50-
COMBINATION NOZZLE, KILN BURNER
GAS INPUT 2734 SCFH; 4|| SCFH AXIAL
2323 SCFH RADIAL
6.6% PRIMARY AIR
WALL TEMPERATURE 1345 C (AIR COOLING)
SECONDARY AIR PREHEAT AS LABELED
1500
1234
O, IN FLUE, %
Figure 29. Normalized NO concentration as a function of O2 in the flue
(excess air) for combination nozzle kiln burner using 6.6$ primary air;
14$ axial-86$ radial injection and 1345°C walls
55
-------
450-
400
350-
300
I
250-
I
-J
200-
150-
100-
50-
COMBINATION NOZZLE, KILN BURNER
GAS INPUT 2714 SCFH; 411 SCFH AXIAL
2303 SCFH RADIAL
3.5 % PRIMARY AIR
WALL TEMPERATURE |320°C (AIR COOLING)
NOZZLE EXIT POSITION
SECONDARY AIR PREHEAT AS LABELED
'457 °C
D
D,
5000
D
V
'8000
0, IN FLUE, %
Figure 30. Normalized NO concentration as a function of O, in the flue
(excess air) for the combination nozzle kiln burner using 3.5% primary air;
14$ axial-86$ radial injection and 1320° C walls;
with nozzle in exit position
56
-------
450
400
350-
I:
ov
300-
250-
200-
150-
100-
50-
COMBINATION NOZZLE, KILN BURNER
GAS INPUT 2687 SCFH; 0 SCFH AXIAL
2687 SCFH RADIAL
3.2 % PRIMARY AIR
WALL TEMPERATURE I305°C (AIR COOLING)
SECONDARY AIR PREHEAT AS LABELED
'2500
1234
0. IN FLUE, %
Figure 31. Normalized NO concentration as a function of O? in the flue
(excess air) for the combination nozzle kiln burner using 3.2% primary ai:
and 0.0$ axial gas injection
57
-------
450
400
350
300-
I
250
§
Nl
-J
^200
O
150-
100-
50-
D
COMBINATION NOZZLE KILN BURNER
GAS INPUT 2659 SCFH; 0 SCFH AXIAL
2659 SCFH RADIAL
WALL TEMPERATURE I340°C (AIR COOLING)
3.5 96 PRIMARY AIR
NOZZLE EXIT POSITION
SECONDARY AIR PREHEAT AS LABELED
P 7500
9100
09 IN FLUE, %
Figure 32. Normalized NO concentration as a function of O2 in the flue
(excess air) for combination nozzle kiln burner in the exit position
and using 0.0^ axial gas injection
58
-------
Similar results to those previously presented were found. At no preheat the
levels were about 90 ppm with a 5$ reduction when compared with baseline
operation and a 19$ reduction when compared with radial injection in the
normal position. At the intermediate preheat (~250°C) level, baseline
operating conditions, total radial injection, and total radial injection in the
exit position produce approximately the same levels of concentration (~135 ppm).
At high preheat (~450°C), the NO concentration of 350 ppm was 21$ higher
than concentrations measured for the normal nozzle position or baseline
operating conditions.
REDUCED FIRING RATE
Figure 33 shows data gathered with the gas input reduced to 1900 SCFH
and 30$ axial gas volume injection. To stabilize the flame, the primary air
had to be increased to 6.2$. As a result of sidewall water cooling, the
average wall temperature decreased to 1023°C. For the 460°C preheat, the
"stoichiometric" NO concentration was less (by approximately a factor of 2)
than that measured for the 2700 SCFH gas input, 1150°C wall temperature,
and 13$ EFGR. The slope of the line for the 1900 SCFH is only 72$ that
for 2700 SCFH and EFGR. Consequently, lower emissions are achieved for
a kiln burner operating in a turndown condition than is achieved with a
higher gas input and EFGR. For intermediate and no preheat, the emission
levels are equivalent for the two operating conditions.
For a gas volume input of 1800 SCFH with only a radial component, the
NO emission levels as a function of excess air are shown in Figure 34. The
slopes of the NO concentration lines for the 1800 SCFH input -was one-half
the slope for the 2700 SCFH input baseline conditions at elevated secondary
air temperature. The ratio of "stoichiometric" NO concentration as a function
of secondary air temperature varied from a low of 1.23 for 245°C to 1.84 for
22°C and 2.98 for 460°C with all temperature cases having the largest NO
levels for the 2700 SCFH firing rate. Thus, a reduction in firing rate to 70$
of the total input resulted in a 58$ reduction in NO emissions or a level of
122 ppm for 3$ excess oxygen and a 460° C secondary air temperature.
59
-------
450-1
400-
350-
E 300-
8:
i
§ 200-
150-
100"
50-
COMBINATION NOZZLE, KILN BURNER
GAS INPUT 1900 SCFH; 570 SCFH AXIAL
1330 SCFH RADIAL
6.2 % PRIMARY AIR
WALL TEMPERATURE i023°c (WATER COOLING)
SECONDARY AIR PREHEAT AS LABELED
456 °C
D
IN FLUE, %
Figure 33. Normalized NO concentration as a function of O2 in the flue
(excess air) for the combination nozzle kiln burner with
reduced gas input (1900 CFH) and 30$ axial gas injection
60
-------
450
400-
350-
300-
250-
-O
200-
150-
100-
50-
COMBINATION NOZZLE KILN BURNER
GAS INPUT 1800 SCFH/ 0 SCFH AXIAL
1800 SCFH RADIAL
4.296 PRIMARY AIR
WALL TEMPERATURE I250°C (AIR COOLING)
SECONDARY AIR PREHEAT AS LABELED
234
, IN FLUE, %
Figure 34. Normalized NO concentration as a function of O2 in the flue.
(excess air) for the combination nozzle kiln burner fired with
1800 SCFH of gas and 0.0$ axial injection
61
-------
Since the operating conditions for the data in Figure 34 were similar to
the baseline conditions (except for the 30$ axial injection), a comparison of
Figures 19 and 34 may yield a rough approximate of the total NO concentration
from combustion of axially injected gas and an approximation of the contribution
from the radially injected gas.
A comparison of "stcichiometric" NO concentrations would indicate half
of the NO concentration was due to axial gas injection at high preheat (~460°C),
37 $ at 250° C and 44$ at ambient. At ambient the major contribution is radially
with the slopes being approximately identical only the stoichiometric difference
of 25 ppm separates the 2700 SCFH, 30$ axial fuel injection from the 1800 total
radial injection. However, at elevated temperatures, the axially gas contri-
bution increases significantly such that at 3 4 excess oxygen for a 250° C
secondary air preheat 42$ of the total NO could have its origin from the axial
gas and for 460°C 58%. Some of the contribution could arise from effective
additional secondary air preheat due to scrubbing of heat off the furnace walls,
resulting in a higher flame temperature.
DIVERGENT FUEL NOZZLE
After the conbination nozzle trials were completed, the kiln burner was
modified to accept the divergent gas nozzle (Figure 2) which permitted in-
jection of the gas at an approximate 45-degree angle relative to the burner
axis. This would simulate a fuel velocity vector similar to the combination
nozzle when operated with 50% of the total fuel volume being injected radially
and 50$ axially. The advantage that the divergent nozzle has is that it
projects all. the gas out into the furnace for combustion rather than having a
large radial portion burned within the burner block. The nozzle is able to
maintain a flame length suitable for the pilot test furnace.
Figure 35 presents NO emissions data versus excess oxygen for baseline
operating conditions with the divergent nozzle. At ambient temperature
secondary air, the 30$ axial injection with the combination nozzle and the
divergent nozzle have nearly identical emission characteristics. At the
intermediate and high secondary air preheat temperatures (250°C and 460°C,
respectively) the NO slope lines for the divergent nozzle are about half
those for the 30$ axial combination nozzle. However, the significant factor
is that the "stoichiometric" NO concentration for the divergent nozzle is
62
-------
450-1
400
350-
I
cr
300-
250-
1
200-
150-
100-
50-
5500
DIVERGENT NOZZLE KILN BURNER
GAS INPUT 2700 SCFH
3.5% PRIMARY AIR
WALL TEMPERATURE i32o°c
(AIR COOLING)
SECONDARY AIR PREHEAT AS
LABELED
234
0 IN FLUE, %
Figure 35. Normalized NO concentration as a function of O2 in the flue
(excess air) for the divergent nozzle kiln burner using
3.5$ primary air; 2700 SCFH gas input and 1320°C walls
63
-------
three times that of the 30$ axial combination nozzle or 303 ppm compared
with 111 ppm. Thus, at 3 $ excess oxygen, with a 460°C secondary air pre-
heat, the NO emission level for the divergent nozzle is 396 ppm, while for
the 30$ axial combination nozzle it is 288 ppm. Consequently, it is again
demonstrated that projecting the flame too far into the furnace, and delaying
bvirnout beyond an optimum point, is an open invitation to increased levels
of NO emissions.
To test the secondary air preheat scrubbing theory, the above tests were
rerun with the walls water-cooled (Figure 36). This resulted in a sidewall
temperature of 1145°C compared to that in the previous trials of 1320°C.
The biggest significant factor was the reduction in the "stoichiometric" NO
concentrations from baseline conditions. For the 460°C preheat there was
a 79$ reduction to 63 ppm; for the 250°C preheat a reduction of 67$, to a
concentration of 59 ppm was observed; and for ambient temperature, a
41 ppm NO level was determined representing a 32$ reduction. Comparing
the data of Figure 36 to the baseline operating data with the 30$ axial com-
bination nozzle shown in Figure 35 indicates that for 460°C preheat and a
3$ level of excess oxygen there is a 54$ reduction in the NO to a level of
134 ppm by using the divergent nozzle with the 1145°C wall temperature.
Thus, if the furnace is maintained below a certain critical temperature,
projecting the flame into the furnace has the desired effect. The combustion
is delayed and products of combustion mix with the fuel and air, localized
combustion intensity is reduced which both help to reduce peak flame
temperatures and lower NO.
As a final reduction measure, the water cooling of the sidewalls was
maintained but the primary air percentage of the total combustion air was
increased to 9.5$. This increase in primary air should result in a larger
volume of cold air (primary air not being preheated) available for combustion
and consequently result in shielding and delaying the mixing of the preheated
secondary air with the fuel. The trial results are presented in Figure 37.
The "stoichiometric" NO concentrations, although slightly higher (~5 ppm),
show no significant change with the additional primary air. The slopes of
the NO concentration lines have decreased by about 25$. Thus, the divergent
64
-------
450
400
350-
•8:
300
250-
-J
200-
150-
100-
50-
DIVERGENT NOZZLE, KILN BUHNER
GAS INPUT 2700 SCFH
3.5 % PRIMARY AIR
WALL TEMPERATURE ii45°c (WATER COOLING)
SECONDARY AIR PREHEAT AS LABELED
Oz IN FLUE %
Figure 36. Normalized NO concentration as a function of O2 in the flue
(excess air) for the divergent nozzle kiln burner using
3.5$ primary air; 2700 SCFH gas and (cooled) 1145°C walls
65
-------
450-1
400-
350-
£300
8:
o-
Q250
Lu
Nj
-J
&200
150-
100-
50-
DIVER6ENT NOZZLE/ KILN BURNER
GAS INPUT 2700 SCFH
9.5 °/b PRIMARY AIR
WALL TEMPERATURE II50°C (WATER COOLING)
SECONDARY AIR PREHEAT AS LABELED
452°C
1234
Oz IN FLUE, %
Figure 37. Normalized NO concentration as a function of O2 in the flue
(excess air) for the divergent nozzle kiln burner operated with
(cooled) 1150°C walls and 9.5^ primary air
66
-------
nozzle with water wall cooling and 9.5$ primary air has an NO emission
level of 109 ppni for 3$ excess oxygen and 460°C secondary air preheat or
a 62$ reduction when compared to baseline operating conditions.
SUMMARY OF KILN BURNER RESULTS
A synopsis of the operational variables studied for the kiln burner and
their test results is presented in Table 3. The technique producing the most
dramatic reductions in NO emission levels wa.s wall temperature, which for
the divergent nozzle resulted in a 66$ decrease in the flue concentration.
With hot furnace walls, external flue gas recirculation produced the largest
reductions in NO with a 49$ decrease. The use of combustion aerodynamics
to recirculate combustion products to the base of the flame (internal flue gas
recirculation) is difficult with the kiln burner since the secondary air has no
control over flame characteristics.
Although variations in NO emissions occur because of changes in the
amount of excess air, NO reduction levels can be established in addition to
their relative effectiveness, by comparing the emission levels at a fixed
level of excess air. The conclusions reached below are based on an excess
air level equivalent to 3 $ oxygen in the flue and a combustion air preheat
temperature of 460°C. For the combination nozzle:
a. 13$ external flue-gas recirculation reduced NO emissions by 49$.
b. Cooling the walls from 1330°C to 1130°C reduced NO emissions
by 29$. However, it is acknowledged that reduced wall tem-
perature may not be a practical technique for some types of
processes or furnaces.
c. Reducing the combustion air preheat from 460°C to 250°C led to
a 50$ reduction in NO emissions. No preheat gave a 68$ reduction.
d. Combining the 1130°C wall temperature with the 13$ external flue
gas recirculation resulted in a 68$ reduction, or an emission level
equivalent to operating with ambient secondary air temperature.
e. Increasing the primary air to 6$ of the total air input yielded a
19$ decrease in NO emissions.
f. Combining the 6$ level of primary air with the 1130°C wall
temperature reduced the flue NO concentration by 24$.
g. Moving the injector to the exit position resulted in an increase in
the NO level of 18$.
67
-------
Table 3. SYNOPSIS OF DATA COLLECTED FOR THE KILN BURNER
Excess Oxygen, $ 3*
Normalized NO, ppm
(Baseline operating conditions; gas input 2700 SCFH; 460°C secondary air
preheat; combination gas nozzle, 810 SCFH axial, 1890 SCFH radial;
normal nozzle position; 1330°C wall temperature; 3.2$ primary air)
Secondary Air Preheat, °C
460 250 22
Baseline Operation 288 143 92
EFGR, 13$ 146 45 28
Wall Temperature, 1130°C 204 94 52
Wall Temperature, 1130°C and
EFGR, 13$ 92 47 22
Primary Air, 6% 233 125 42
Primary Air, 6$ and Wall
Temperature, 1130°C 219 105 68
Exit Nozzle Position 339 135 62
Axial Gas, 14$ 303 173 100
Axial Gas, 14$ and Exit Nozzle Position 323 136 73
Axial Gas, 14$ and Primary Air, 6$
Radial Gas, 100$ 291 133 108
Radial Gas, 100$ and Exit Nozzle
Position 352 136 87
Gas Input, 1900 SCFH 89 44 29
Gas Input, SCFH and Radial Gas 100% 122 83 62
xf
NO concentrations at 1$ excess oxygen have been omitted. Many operating
conditions produced CO concentrations above 500 ppm making these con-
ditions undesirable despite a low NO level.
Divergent Nozzle 396 231 104
Divergent Nozzle and Wall
Temperature, 1100°C 134 115 67
Divergent Nozzle, Wall Temperature,
1100°C and Primary Air, 9.5$ 109 100 69
68
-------
h. Changing the ratio of axial gas to total gas to 14$ (from 30$)
resulted in a 5* increase in NO emissions. By increasing the
primary air to 6% .(irora 3.5$), the level was brought back to
that of baseline operation conditions.
i. With 100$ of the gas being injected radially the emission levels
were equivalent to baseline. Moving the nozzle into the exit
position, however, resulted in a. 22% increase in emissions.
j. Reducing the firing rate from 2700 SCFH to 1900 SCFH resulted
in a 69$ reduction in NO emissions.
For the divergent nozzle:
a. The divergent nozzle increased NO emissions by 38$ compared
to the combination nozzle. However, reducing the wall tempera-
ture resulted in a 53$ reduction of NO emissions relative to
baseline operations. Using the reduced wall temperature and 6%
primary air produced an overall reduction of 62%.
It can be concluded for the kiln burner that there is an optimum flame
shape and length that will produce minimum NO emissions. Too short a
flame will produce high combustion intensity within the burner block thus
causing high NO emissions. Projecting the flame to far down the furnace
will result in additional preheating of the secondary combustion air by
"scrubbing" heat from the furnace walls resulting in both higher flame
temperatures and increased NO emissions. An effective procedure is to
minimize wall temperature where practical (from data this should be below
1100°C) and maximize the volume and velocity of axial fuel (staying within
limits of stable combustion and low flue CO). Short of this, increasing the
volume of "cold" primary air is the easiesst technique to follow.
69
-------
BAFFLE BURNER
Figure 4 illustrates the design of the baffle burner. The gas nozzle lies
parallel to and along the axis of the burner. It is inserted into the ceramic
baffle, thus ensuring that the gas enters parallel to the axis of the baffle
burner. Combustion air enters perpendicular to the axis and passes through
the six ports in the baffle, which impart a swirl to the air in some designs.
Two of these baffle ports are shown in Figure 4 with their axes parallel to
the axis of the burner. Air exiting from the ports (as illustrated) would have
only an axial velocity component, resulting in a "long" flame. To shorten
the flame length, a tangential-flow component must be added to the combus-
tion air velocity. This is accomplished by using a baffle where the
combustion-air ports are rotated relative to the axis of the burner. For
the "intermediate" flame length baffle (IFLB), the rotation orientation of the
ports is 15 degrees and for the "short" flame length (SFLB) it is 25 degrees.
IFLB BURNER
Standard Conditions (IFLB)
Figure 38 presents normalized nitrogen oxide (NO) versus excess oxygen
(O2) test data for burner conditions typical of an industrial situation (standard
gas nozzle, baffle nozzle position, and 4 degree burner-block angle). Mea-
sured CO levels greater than 500 ppm are listed by concentration next to their
corresponding data point. Photographs illustrating flame geometry and flame
luminosity under these industrial conditions are shown in Figures 39 and 40.
The standard gas nozzle is a 2-inch pipe that rests within a centering
tube mounted between the baffle and the burner housing. Gas nozzle positions
are illustrated in Figure 41. The gas nozzle position is denoted as the baffle
position when the nozzle is even with the burner-block side of the baffle.
Other nozzle positions investigated were the throat position ( gas nozzle mid-
way between the baffle and the front wall of the furnace) and the exit position
( gas nozzle even with the front wall of the furnace) . For the IFLB, the exit
gas nozzle position produced unstable combustion and therefore is not
included in the test results.
70
-------
600-1
Gas Input 3070 ft 2O05 6CFH
Cos Nozzle Baffle Position
Won Temperature 1435* c
4 Burner Block Anglt
500-
400'
E
ex
a.
O
2
•a
a>
N
E
b
z
300-
200-
100-
A. 1,700
13,OOO
2p£%r
o
1 6,000
1
1
-~o"~^
D — S &
o— o
1
2
A
O
1
3
_ — —
. i A n
Standard
Standard
1
4
. 22° C
Legend
Nozzle,
Nozzle.
1
5
22°
-(
Gos
Gas
•^v
^
Input
Input
. 1
6
3000
1 OOO
CFH
CFH
IN FLUE,, %
Figure 38. Normalized NO concentration as a function of flue O2 for
the IFLB burner with a standard gas nozzle at gas inputs
of 3070 and 2005 SCFH
71
-------
Figure 39. IFLB burner with standard fuel nozzle
Figure 40. IFLB burner with divergent nozzle
72
-------
Figure 41. Nozzle positions tested for the baffle burner
73
-------
In-the-flame surveys were conducted during the baffle burner trials.
These surveys included flow direction, temperature and gas analysis. The
details of this survey work along with complete in-the-flame data and a
comprehensive listing of the input/output data are presented in Volume II.
However, it is beneficial in understanding the interpretation of the input/
output data to take a brief look at some in-the-flame data.
Figure 42 presents some gas species and temperature data collected
along the baffle burner axis for standard operating conditions. The flow
direction data shows the profile of a type I flame. The axial momentum of
the fuel jet is less than that of the combustion air. At the burner block exit
there is a methane concentration of less than 2^. The CO concentration
decreases from the burner block exit to the end of the furnace, with the
maximum measured concentration being 9.2$. The temperature at the
burner block is quite high at 1610°C. The NO concentration is 120 ppm or
24^ of the 493 ppm flue concentration.
At a gas input of 3000 SCFH, the relationship between NO and excess
O2 became increasingly nonlinear as a function of temperature, as shown
in Figure 43. This indicates that in the range of excess O2 investigated, the
higher the preheat, the more closely the NO-versus-O2 relationship
characterizes that of a premix flame.
For the second set of curves in Figure 43, burner conditions were the
same, but the gas input was reduced to 2000 SCFH. In an attempt to main-
tain the same level of bulk NO formation, the wall temperature was held at
1420°C, compared with 1435°C for a gas input of 3000 SCFH. These curves
show the same linear relationship between NO and excess O2 that was ob-
served for the kiln burner. This linear relationship is characteristic of a
diffusion flame.
External Flue Gas Recirculation (EFGR) for IFLB
As a base-line control case, the effect of external flue gas recirculation
was tested. The percentage of FGR is determined by using the relationship —
EFGR (SCFH)
X 1UU
Secondary air (SCFH) + Fuel (SCFH)
74
-------
lo
T°C
*O^/">
I8OO —
1600-
I40C-
\200-
1000-
80C-
xnn .
V
N0 PF
C f~\f^\
3UU —
400-
300-
'200-
100-
\
n --
A
>m CC%
—
9-
8-
7-
6-
5-
"
4-
3-
2-
1 -
r, -•
0
; c-H4
*
18 -
16 -
14 -
12 -
!C -
8 -
6 -
4 -
2 -
r
o/
~A\ ,o-^^~
A ^ ^'
\ S*
\ x^
/
' /
/\..
V
/
/ v
/
/ A
/ \
^ \x
/ . v\
i ^\
i
i
"""-o
50
150
DISTANCE FROM BURNER - CM
Figure 4Z. In-the-flame profiles along the axis of a baffle burner
for typical operating conditions
-------
600
500-
400-
£
CL
0.
3OO
N
O
£
t_
O
z
20O~
100-
O
1,700
Gas input 3070 SCFH
Gos Nozzle Baffle Position
Wall Temperature 139O c
4 Burner Block Anole
-°I2,OOO
I4,OOO
1
1
o- — o
1
2
°2
_
1 1
3 4
IN FLUE, %
— o-^p * f
1 1
5 6
FOR
Figure 43. Normalized NO concentration as a function of flue O2 for
the IFLB burner with a standard gas nozzle
and 154 and 30$ FGR
76
-------
By using the data presented in Figure 43, the NO concentration measured
for standard burner operating conditions with 462°C secondary air preheat
can be compared with the concentration at similar burner conditions but with
the addition of 15* and 30^ EFGR to the secondary air. During preheat,
both the flue gas and secondary air were blended and reached the same final
preheat temperature, which in this particular test was 460°C. At 3$ excess
O2, a 15^ FGR level reduced the normalized NO from 575 to 150 ppm,
whereas 30$ FGR reduced the measured NO concentration to 50 ppm.
Wall Temperature (IFL.B)
The data for Figure 44 were1 collected with the burner operating in the
normal industrial mode; however, the wall temperature was decreased from
1435° to 965 °C by using water-cooling tubes within the furnace sidewalls.
This reduction in Avail temperature resulted in approximately a factor
of two decrease in the level of NO emissions. It is not possible to determine
from the data what the magnitude in reduction of the prompt or bulk NO levels
was. This in bulk NO formation occurs because of the increased heat removal
in the post flame region and secondary recirculation zones. When these
cooler recirculation zone products are entrained by the combustion air jet,
the peak flame temperature is lowered resulting in a decrease in the prompt
NO level.
Although it is not possible to operate with reduced wall temperatures
(below 1200°C) in all industrial applications, these data demonstrate its
effectiveness as an NO control technique. Suggestions for achieving lower
operating wall temperature are monitoring fuel input, increased product
load and/or secondary furnace cooling system.
Nozzle Type and Position (IFLB)
Figure 45 presents NO concentrations measured under standard burner
conditions but with a radial gas injector. This injector forces the gas to
enter the burner block radially to the axis of the burner, which causes a
flame flow pattern categorized as type III. The combination gas nozzle shown
in Figure 3, operating with radial injection only, was used.
77
-------
300-
25O-
E
Q.
CX
*
Z 200-1
4)
N
150-
o
Z
100-
50-
Gos Input 3070 SCFH
Gas Nozzle Baffle Position
Wall Temperature 965° C
4° Burner Block Angle
Air Preheat AS Labeled
O
232 C
O Standard Nozzle, Baffle Position
Wall Temperature 965° C
1 1
1 2
°2
IN
1
3
FLUE,*
1
4
•
1
5
1
6
Figure 44. Normalized NO concentration as a function of flue O2 for
the IFLB burner with a standard gas nozzle
at a wall temperature of 965°C
78
-------
700-1
600-
500'
e
a.
CL
'
O
Z
4OO-
-------
The radial nozzle in the baffle position produced more NO than the
standard nozzle because of the increased rate of heat release, coupled with
the inability of the burner block to be used as a heat sink. However, when
the radial nozzle was moved into the throat position, there was a dramatic
drop in the measured NO concentration (approximately a factor of 2-1/2).
This decrease could be due to the combustion occurring outside the burner
block, which allows a larger mass exchange to occur among the fuel, the
internal recirculation zone, the secondary air, and the external recirculation
zone, thus lowering the flame temperature.
In an attempt to further slow the rate of heat release, half the gas was
introduced radially while the other half was introduced axially (at a high
velocity of approximately 300 ft/s). Again, data were collected for the half
axial-half radial injection in both the baffle and throat positions. These
results are presented in Figure 46. The axial gas injection had the desired
effect: it decreased the NO concentration relative to all-radial injection by
projecting the flame further into the furnace and slowing the rate of heat
release. Again, moving the nozzle position into the throat of the burner
block further reduced the NO concentrations.
Figure 47 is a composite plot showing all significant reductions in NO
as a function of nozzle type and/or nozzle position. All these data were col-
lected with a secondary air preheat at 460°C. The base-line conditions were
those of the standard nozzle in the baffle position; this burner configuration
produced the highest NO concentrations. Moving the standard nozzle into the
throat position did not change the shape of the NO-versus-O2 relationship,
indicating that the basic mixing patterns were not dramatically altered. How-
ever, it did decrease the NO concentration to approximately 125 ppm. With
a divergent gas nozzle (Figure 2) similar to the one used in the kiln burner,
an additional decrease to 100 ppm NO resulted. The differences in NO con-
centration produced with the divergent nozzle and the standard nozzle in the
throat position decreased as a function of excess O2, until at 6^ excess O2,
the NO concentrations were equal. The divergent nozzle produced a type III
flame. The decrease in NO occurred because of the additional entrainment
of combustion products from the primary and secondary recirculation zones.
Although the data are not plotted, test results with the divergent nozzle in
80
-------
600-
500-1
400-
£
Q.
ex
N
0
200-
100-^
,700
Gas Input 3101 SCFH
1511 Axial, I59O Radial
Gas Nozzle Baffle & Throat Position
Wall Temperature I39O° C
4° Burner Block Angle
Legend
A Standard Nozzle, Baffle Position
O Half Radiol Axial Nozzle. Baffle Position
D Half Radial Axial Nozzle, Ttiroat Position
I
4
I
5
I
6
IN FLUE,%
Figure 46. Normalized NO concentration as a function of flue O2
for the IFLB burner with a combination gas nozzle
and axial and radial injection
81
-------
600-1
500-
400-
E
a.
o.
«
O
z
•o
0)
N
O
E
300-
200-
100-
Gos input 3070 SCFH
Secondory Air Preheot 460° C
4° Burner Block Angle
Legend
O Standard Nozzle, Baffle fbsltlon
VStandard Nozzle, Throat Position
A Divergent Nozzle, Baffle Position
D Axial Nozzle, Baffle Position
OStandard Nozzle, Bdffle Position, 15% FOR
°
3
N FLUE,
Figure 47. Normalized NO concentration as a function of flue O2 for the
IFLB burner with the various gas nozzles in different positions
82
-------
the throat position paralleled those with the kiln burner; i.e., the NO was
increased by moving the gas injector toward the combustion chamber. The
difference in NO concentration as a function of nozzle position, however, is
very small.
The most dramatic decrease in NO with burner alterations occurred
when the gas velocity was increased by a factor of 16. This was accom-
plished by using all-axial gas injection, but decreasing the injector opening
from the standard 2 inch to 0.5 inch. As shown in Figure 47> at less than
24 excess O2, the divergent nozzle produced the lowest NO emissions; how-
ever, above that level of excess O2, the high-velocity axial nozzle produced
the lowest NO emissions. At 3$ excess O2 (11.4^ excess air, a typical
industrial level), comparing the normal burner output of 515 ppm with the
high-velocity gas-injection burner output of 250 ppm, NO emissions were
decreased by 325 ppm through a simple, low-cost alteration of the gas in-
jector. Note that external FGR still remains the best depressant of NO
formation that was tested. At 15$ FGR, the NO level is 150 ppm with the
standard baffle-and-nozzle configuration, which represents a total decrease
of 425 ppm.
The control condition in-the-flame data was collected for the high
momentum axial nozzle. Figure 48 presents some of the gas species and
temperatures measured along the burner axis. As with the standard operating
conditions the flow profile is a type I. The difference between the profiles is
that now the primary jet axial momentum is greater than the axial momentum
of the combustion air. This is reflected in the high concentration of methane
(154), measured at the burner block exit. The temperature at the burner
block exit is only 1265°C or about 350°C cooler than that from standard
operating conditions. The NO concentration at the burner block exit is
21 ppm or 8% of the 254 ppm flue concentration.
Burner-Block Angle (IFLB)
To test the influence of the primary recirculation zones on the level of
NO emissions, the burner-block angle was increased from 4 to 8 degrees.
The primary recirculation zones contain not only combustion products but
also fuel and air. The entrainment of these gases by the primary and
secondary jets should help to depress the flame temperature; however, they
83
-------
o
v
A O
oo
T'C - NO ppm C0% CHa%
HOC
tcoo-
I40O-
1200-
1000-
800-
600-
200-
180-
160-
140-
120-
100-
80
6O
40-
20-
o-
10-, 20
50 100
DISTANCE FROM BURNER - CM
Figure 48. In-the-flame profiles for the axial fired baffle burner
in the controlled operating condition
15
-------
are not as desirable as the gases in the secondary recirculation zone
because they contain both fuel and air. The angular difference between the
IFLB (the air ports are angled at 15 degrees relative to the centerline of the
baffle) and the burner block (air entry being an 8-degree divergent cone) is
7 degrees. Because the angle of the ports is greater than that of the burner
block, the flow will not separate from the contour of the block. The IFLB
was tested with the standard, divergent, and high-velocity axial gas nozzles
with an air preheat temperature of approximately 460° C. The results are
shown in Figure 49.
Table 4 compares the NO levels for integer values of the percentage of
excess O2 in the flue versus nozzle type between the 4 and 8-degree burner
Table 4. NORMALIZED NO AS A FUNCTION OF NOZZLE TYPE,
BLOCK ANGLE, AND EXCESS Oj, FOR THE IFLB BURNER
WITH AN AIR PREHEAT TEMPERATURE OF 46o°C
Block Angle,
Nozzle Type degrees
Standard 4
8
Divergent 4
8
Axial 4
blocks. For the standard and divergent gas nozzles at low levels of excess
O2 (below 3^), the emission levels from the two burner blocks are compar-
able. However, as the excess air is increased, the emission levels from
the 8-degree block become larger than those from the 4-degree block. Use
of the high-velocity (611 ft/s) axial nozzle with the 8-degree block at normal
operating levels of excess O2 (below 3^) resulted in a 20$ reduction in NO
emissions. Thus the desired mass exchange among the primary jet,
secondary jet, and the internal recirculation zones occurs only for the high-
velocity axial nozzle. In this case of the standard nozzle (38 ft/s gas velocity)
and the divergent nozzle (654 ft/s gas velocity), the increased burner volume
resulted in high emissions. This increase occurred because a) the rate of
heat release resulting from a high mass exchange between the primary and
85
1
390
408
197
204
227
174
2
487
492
273
272
255
196
3
ppm
577
564
313
360
280
238
4
567
612
340
428
300
278
5
542
636
360
480
313
313
-------
700-
Gas Input 2298 SCFH
Gcs Noz7le Baffle Position
600-
500"
E
ex
a.
»
O
z
•o
a>
N
o
E
400
300-
200-
100"
Legend
O Standard Nozzle
A Divergent Nozzle
O Axial Nozzle
— 1
1
1
2
1
3
I
1
5
1
6
IN FLUE,
Figure 49. Normalized NO concentration as a function of flue O2 for
the IFLB burner with standard, divergent, and axial gas nozzles
86
-------
secondary jets improved more than the mass exchange among the primary
jet, the secondary jet, and the internal recirculation zones or b) the volume
of the secondary recirculation zone was decreased.
SFLB BURNER
Air Velocity (SFLB)
Initial tests with the SFLB burner were conducted by using a high-
velocity (126 ft/s) baffle and a low-velocity (80 ft/s) baffle. (Air injection
ports are angled at 25 degrees relative to the centerline of the baffle.)
Because of the extreme pressure drop associated with the high-velocity
baffle (approximately 22 in. H2O with an air flow rate of 38,000 SCFH at
450°C), the tests were conducted with a gas load of 2000 SCFH. The 8-degree
burner-block angle used in these tests is the manufacturer's recommended
angle for industrial applications.
Figure 50 presents the test results as normalized NO emissions plotted
against the measured percentage of O2 in the flue. The high-velocity baffle
produced consistently lower emission levels than the low-velocity baffle as
a result of the larger mass exchange between the secondary recirculation
zone and the secondary jet due to the increased velocity gradient. Thus,
the higher the velocity of the secondary jet, the lower the NO emission
levels. The magnitude of the reduction in NO depends not only on the velocity
gradient across the shear layer separating the secondary jet from the
secondary recirculation zone, but also on the volume and temperature of the
secondary recirculation zone. The cost involved for this method of NO
reduction is additional fan horsepower required to overcome the increased
pressure drop in the burner.
Wall Temperature (SFLB)
Figure 51 illustrates the effect of wall temperature on the NO emission
levels of the SFLB as a function of secondary air preheat and excess air.
Conclusions similar to those for the IFLB can be drawn for the SFLB: A
wall temperature as low as possible must be maintained through increased
wall cooling, lowered gas input, or furnace design.
87
-------
600-
500-
o. 400-
o.
»
O
:r 300 H
o
o
z
200-
IOO-
Gas input 2020 SCFH
Gas Nozzle Baffle Position
Wall Temperature 1295° C
Air Preheat As Labeled
8° Burner Block Angle
Legend
130/XX) o LOW Velocity (60 fps) Short Flame Bofflt
A High Velocity (125 fps) Short Flam« Baffl*
IT]]
2345
C IN FLUE, %
i
6
Figure 50. Normalized NO concentration as a function of flue O2 for the
SFLB burner with a standard gas nozzle
88
-------
E
Q.
Q.
800-t
700-
600-
5OO-
N
£ 400H
o.
3OO-
200-
IOO-
3079 SCFH. Axial
Gas input
Gas Nozzle Baffle
Wall Temperature 1050
Air Preheat As Labeled
8° Burner Block Aoale
Position
° a
1450 C
30,000
20,000^
ll,OOO
452° c
-- Legend
°
O 1460 c Wall Temperature
A (050° C Wall Temperature
42O° c
210° C
•— O"""
^^fr^**L
O*"""'^
A A--
-A-
22° C
___- — O—
&
212° C
"~A 22° C
~T
4
—i
6
02 IN FLUE ,96
Figure 51. Normalized NO concentration as a function of flue
SFLB burner with a standard gas nozzle at wall
temperatures of 1450° and 1050° C
for the
89
-------
External FOR (SFLB)
To determine •whether combustion aerodynamics can be used to recir-
culate combustion products to the base of the flame as an emission control
technique, we simulated the idealized case (in which the combustion air and
products are thoroughly mixed before ignition) by mixing flue gases with the
combustion air outside the burner. Figure 52 compares the NO levels as a
function of the percentage of C>2 in the flue for 0$ , 15$, and 25$ FGR. Again,
as demonstrated for the kiln burner and the IFLB, external FGR is an
extremely effective method of controlling NO emissions from the SFLB.
Nozzle Type (SFLB)
To determine the type of injector that would reduce NO emissions, a
series of trials were conducted by using different methods of gas injection;
the results are shown in Figure 53. The NO-versus-excess O2 relationship
for the standard nozzle (injection method recommended by manufacturer;
2-inch stainless-steel pipe with an inlet velocity of 38 ft/s and a gas input
rate of 3000 SCFH)is shown as a reference, so that the effect of the gas in-
jector type on the NO emission levels can be evaluated. Visual observations
were made to determine the flame length for each gas nozzle tested. This
allowed a first-order evaluation of alterations that can be made to obtain the
flame shape desired by the manufacturer.
The visual flame length observed with the standard gas nozzle with the
SFLB was 103 cm. The high-velocity (1043 ft/s) radial injector increased
the NO emissions substantially above those measured for the standard nozzle.
The shape of the emission curve closely resembled that of a premixed flame;
peak emissions occurred at 1.5$ excess O2, The flame was invisible. Re-
ducing the radial velocity to 532 ft/s resulted in a shift of the maximum NO
emission to an excess C^ level of 2.3$. Because of the reduction in the
radial injection velocity, the magnitude of the peak NO concentration was
reduced 20$. The visual flame length was 72 cm, still 31 cm shorter than
that desired. The divergent nozzle, which was the first altered gas injector
tested with the SFLB, yielded lower emission levels than the standard nozzle.
The gas velocity from the divergent nozzle was 654 ft/s (at a 3000 SCFH gas
load); however, a type III flame resulted because of the wake generated by
90
-------
800-
700-
600-
E
a.
ex
O SCO-
N
-400
o
e
o
300-
Gas input 3O7O SCFH, Axial
Gas Nozzle Baffle Position
Wall Temperature I36O° C
8° Burner Block Angle
0% FOR
30,000
»%/w
100-
_ O
6,OOO — -O'~~~~~O
O^-"-"^0
•^^^ ^*\ ^_^H^«- f*\ i
o-o— °"
1 1 1 1
1234
15% FOR
O
2596 FOR
o-
1 1
5 6
IN FLUE, %
Figure 52. Normalized NO concentration as a function of flue O2 for the
SFLB burner with a standard gas nozzle and 15$ and 25% FGR
91
-------
E
ex
Q.
N
e
O
1000
900-
800-
7OO-
600-
500-
4OO-
3OO-
2OO~
100-
Gas Input 3062 8CFH
Gas Norzle Baffle Position
Wall Temperature 1420° C
8° Burner Block Angle
.0
V Radial Nozzle (1043 fps Gas Velocity)
^ Radial Nozzle (532 fps Gas Velocity)
O Standard Nozzle
O Divergent Nozzle
O Axial Nozzle
O Standard Nozzl» 15% FOR
.0^o- o-
1
1
1
2
1
3
4 i I
IN FLUE ,
Figure 53. Normalized NO concentration as a function of flue O2 for the
SFLB burner with the various gas nozzles
92
-------
the divergent cone in the nozzle. Despite the high inlet velocity, the flame
length was 113 cm, which compares favorably with the desired length of
103 cm.
The injection method that resulted in the lowest emissions was the high-
velocity axial nozzle (611 ft/s and a 3000 SCFH gas load). This nozzle
produces a type II flame, even with the high-swirl SFLB, because the velocity
of the gas is large enough to split the primary recirculation zone. Although
the rate of entrainment per unit area between the primary jet and the sur-
rounding flow zones increased because of the larger velocity gradient when
compared with that from the standard nozzle, the flame length also increased,
to 286 cm. The increased flame length was a result of a lower rate of mass
exchange between the primary and secondary jets. This reduced mass ex-
change resulted from a decrease in the area of the shear layer caused by
the smaller primary jet. Thus, by delaying the mixing between primary and
secondary jets, not only was the heat release rate of the flame slowed, but
it also approached the ideal of the EFGR by allowing more time for the
secondary jet and the combustion products in the secondary recirculation
zone to mix.
Burner-Block Angle (SFLB)
Figure 54 gives NO data from the SFLB trial with standard, divergent,
and high-velocity axial nozzles and a 16-degree burner block. Obviously,
the combustion aerodynamics were changed drastically as a function of
excess air. There is no one general interaction theory that explains these
curves in their entirety. Therefore, 1$ level of excess O2 was selected for
analysis. This level was chosen because it consistently produced the lowest
levels of NO, with concentrations below 50 pprn.
Table 5 lists the nozzle type as a function of burner-block angle for both
the SFLB and the IFLB. The inlet gas velocity is listed in parenthesis beside
the nozzle type. For 14$ excess air (3$ O2), the velocity is 120 ft/s. Note
that the air velocity increases with increasing excess air. The arrows in-
dicate the direction of increasing NO concentration. It is known from flame-
length observations that the standard and divergent nozzles have similar
rates of mass exchange between flow zones, whereas the axial nozzle has a
93
-------
60CH
500-
E
Q.
d
400-
300H
N
O
I
200-
100-
Gos input 2953 SCFH
Gas Nozzle Baffle Position
Wall Temperature 1460° C
16° Burner Block Angle
— Legend
o Standard Nozzl*
A Divergent Nozzle
a Axial Nozzle
234
IN FLUE, %
T
5
I
6
Figure 54. Normalized NO concentration as a function of flue O2 for the
SFLB burner with standard, divergent, and axial gas nozzles
94
-------
Table 5. NORMALIZED NO CONCENTRATION AT 1 4 EXCESS O2 WITH
AN AIR PREHEAT TEMPERATURE OF 460°C AS A FUNCTION OF
BAFFLE TYPE, GAS NOZZLE TYPE, AND BURNER-BLOCK ANGLE
NO Concentrations
IFLB SFLB
Block Angle, degrees
16 8 16
_ Nozzle Type --- PPm
Radial (1043 ft/s) -- -- -- 950
Radial (532 ft/s) 590 -- -- 700
Standard (38 ft/s) 375- 400- 470 490- 240
Divergent (654 ft/s) 175- 200- 235 440- 210
Axial (611 ft/s) 215- 175- 200 230- 310
Standard, 15$ FGR 120 -- -- 110
Qz)
slower exchange rate. It is not surprising that the standard and divergent
nozzle show similar trends in NO formation as a function of burner -block
angle. For the IFLB, the NO concentration increases with burner-block
angle; for the SFLB, it decreases with an increasing angle.
For the IFLB, the mixing rate was increased between the primary and
secondary jets more than the size, shape, and mixing rate of the primary
recirculation zone with the primary and secondary jets. For the SFLB, the
mass exchange between the primary recirculation zone and the primary and
secondary jets was improved, which resulted in approximately a 55$ reduc-
tion in NO emissions. The high-velocity axial nozzle showed almost the
opposite effect. For the SFLB, there was a 30$ increase in the emission
level as the block angle was increased. Thus, for high swirl intensity and
high axial velocity, increasing the burner-block angle resulted in a larger
primary- secondary jet mass exchange. Investigations of the IFLB with the
16-degree block to study flow separation between the secondary jet and the
block gave negative results. Therefore, it wa;3 assumed that it (flow
separation) was also not present for the SFLB and the 16-degree block. The
results for the IFLB and the high-- velocity axial nozzle were somewhat mixed.
There was a decrease in the NO level with a block angle increase from 4 to
8 degrees followed by an increase in the NO level with an angle increase from
95
-------
8 to 16 degrees. The optimum burner-block angle for producing minimum
NO emissions with the axial nozzle was approximately 10 degrees.
Wall Temperature (SFLB) Sensitivity
Because of the observed sensitivity of NO to wall temperature during
these tests, NO data was taken during a warm-up cycle. The standard
nozzle was fired with 3000 SCFH of gas at 3 ^ excess O2, 460°C air preheat,
and the SFLB. The results are shown in Figure 55. The dramatic increase
in the level of NO emissions above 1300°C indicates that an accurate control
of wall temperature is required to produce a consistent set of experimental
data. It also confirms that the furnace should be operated with as low a
wall temperature as possible. It may not be possible to operate all industrial
processes with a reduced wall temperature; however, because of its effec-
tiveness as a control technique, operating wall temperature should be
maintained as low as is practical.
FURNACE GEOMETRY
The IFLB burner was mounted on IGT's cylindrical (tunnel) test furnace,
which has a cross-sectional area of 1.17 sq m (12.6 sq ft) and a volume of
6.4 cu m (226 cu ft). The following operating parameters were investigated:
secondary combustion air preheat, method of gas injection, and position of
gas injector.
Figure 56 shows the emissions data collected from the IFLB with the
standard gas injector (2-inch stainless-steel tube), a 2000 SCFH gas input,
and nonpreheated and preheated secondary combustion air. To determine the
influence of furnace geometry on NO emissions, Figure 56 compares these
data with similar data collected during a test on the larger-diameter furnace.
The cross-sectional area of this rectangular furnace is 2.3 sq m (25 sq ft)
and the volume is 10.6 cu m (375 sq ft). With the exception of secondary air
preheat temperature (350°C for the cylindrical furnace, compared with454°C
for the rectangular furnace), all other burner and furnace operating conditions
(gas input, gas nozzle position, burner-block angle, and wall temperature) were
identical. For both conditions of secondary air preheat, the cylindrical furnace
had the higher levels of emissions. The area ratio between the burner-block
opening and the burner wall is 35.9 for the rectangular furnace and 18.2 for
96
-------
70 a
600-
E
a,
a,
50°
o
E
i_
o
z
400-
Gas Input 3O08 SCFH , AXial
Gas Nozzle Baffle Position
3 96 E x c es s O xy gen
460° C Air Preheat
8° Burner Block Angle
/
o
1000
100
I20O
1300 1400 I5OO
Wo 11 Temperature, C
Figure 55. Normalized NO concentration as a function of wall temperature
for the SFL.B burner with a standard gas nozzle
97
-------
800
700
60O
g 500-
4001
Nl
I
O
300-
200-
100-
Gas :nput 1958 SCFH
Gas Nozzle Baffle Position
Wall Temperature 1360° C
Secondary Air Preheat A3 Labeled
4° Burner Block Anflk
— Legend —
o Cylindrical
O Cylindrical
A Rectangular
v Rectangular
-A A 27T
—r~
2
1
IN FLUE . %
Figure 56. Normalized NO concentration as a function of flue O2 for the
IFLB burner with a low-velocity gas nozzle
98
-------
the cylindrical furnace. These ratios reflect the relative sizes of the
secondary recirculation zones. The primary reason for the difference in
measured NO concentrations between the two combustion chambers is the
unequal secondary recirculation zone volume.
Nozzle Type Variations
Additional tests were conducted to investigate the influence of gas in-
jection on total NO emissions. In these trials, both a high-velocity (127 m/s,
407 ft/s) gas nozzle and a divergent gas nozzle (injection velocity 135 m/s,
436 ft/s) were used. Both injectors have only an axial velocity component.
Figure 57 compares the normalized NO emissions as a function of excess O2
from these gas nozzles. The low-velocity nozzle produces less NO, by a
factor of 2, than the high-velocity nozzle. This result contradicts the data
collected from the rectangular furnace, which are presented in Figure 58.
Although the rectangular furnace has a higher gas input (2998 SCFH) and a
higher secondary air temperature (450°C), the cylindrical furnace had the
higher emission levels for a given gas nozzle. This occurs because NO
production in gas flames is entirely thermal in origin. Thus, any change in
the combustion chamber or fuel-air mixing that results in lower peak and
average flame temperatures reduces the rate of NO production. Conditions
that should reduce NO emissions include decreasing the rate of heat release,
increasing the mass exchange between the flame zone and secondary
recirculation zone, and reducing the wall temperature.
During previous trials with the rectangular furnace, the high-velocity gas
nozzle always produced less NO emissions than nozzles with lower injection
velocities. This result was explained by the different fuel-air mixing rates.
An increase in injection velocity was achieved by decreasing the cross-
sectional area of the injector, which caused a decrease in the fuel-air mixing
rate and permitted an increase in recirculation and combustion zone mass
exchange. The combustion zone is blended with combustion products, which
result in lower flame temperatures. However,, for the cylindrical furnace,
the high-velocity gas nozzle causes higher NO emissions. This means that
even though the same method of gas injection was used in each furnace, the
thermal history of the two flames was different. The major reason for a
higher flame temperature in the cylindrical furnace would be a smaller mass
99
-------
BOOn
1200-
iioo-
1000-
e 900
o.
o.
O
Z
Nl
O
O
z
800'
7OO
600'
5OO
Gas input 1958 SCFH
Gas Nozzle Baffle Position
Wall Temperature 1360° C
Secondary Air Preheat 350° C
4° Burner Block Angle
Legend
O Low Momentum Nozzle
A High Momentum Nozzle
^ Divergent Nozzle
234
Cfe IN FLUE. %
Figure 57. Normalized NO concentration as a function of flue O2 for the
IFLB burner with low-velocity, high-velocity,
and divergent gas nozzles
100
-------
600-1
500-
400-
E
a.
a.
XD
OJ
N
O
e
O
300-
200
100-
Gas input
Secondary
4 Burner
3070 SCFH
Air Preheat
Block Angle
450° C
-Legend
OStandoffl Nozzle, Baffle Position
ADivergent Nozzle, Battle Position
p Axial Nozzle, Battle Position
4
IN FLUE,
Figure 58. Normalized NO concentration as a function of flue O2 for the
IFLB burner with standard, divergent, and axial gas nozzles
101
-------
exchange between the secondary recirculation and the combustion zones. The
smaller mass exchange occurred because of a decrease in the size of the
secondary recirculation zones, which, in turn, resulted from a smaller
combustor cross-sectional area.
The divergent nozzle injects the gas radially afc a 45-degree angle to the
burner axis. Comparison of Figures 57 and 58 shows that the cylindrical
furnace has the higher levels of emissions for the divergent nozzle. Argu-
ments similar to those presented for high-velocity nozzle could also be made
for the divergent nozzle. However, for both combustion chambers, the
divergent nozzle has lower emissions than the high-velocity nozzle. The
diverging gas injection causes a primary recirculation zone, which allows a
larger dilution of the combustion zone with combustion products than an
axial injector. This increased dilution leads to lower flame temperatures
and reduced emissions.
NATURAL-GAS, LOW-NO BURNER I
x
IGT has been commissioned by three gas utilities (Consumer's Gas Co.,
Consolidated Natural Gas Co., and Southern California Gas Co.) to develop a
burner with low NO emission levels relative to the industrial burner it was
intended to replace and with a similar flame geometry and luminosity. The
initial design guidelines for this burner development work were extracted
from the data collected in this study. After fabrication, the low-NO burner
' x
I (LNO-I) was mounted on the cylindrical furnace for NO emission level
testing.
Standard Operating Conditions
The initial set of trials was conducted by using 25^ primary air. This
percentage is determined by ratioing the volume of primary air to total air
needed for stoichiometric combustion and multiplying it by 100. Figure 59
shows the results of these tests as normalized NO plotted against excess O2.
These data were collected by premixing the primary air and fuel, which gives
rise to a clear flame. For comparison, Figure 59 also shows the emissions
curve for the IFLB burner with a standard nozzle. The LNO-I burner was
designed to replace burners in the IFLB burner category. Comparison of
the emission levels of these burners at 350°C secondary air preheat shows
that the LNO-I burner was extremely effective in reducing NO emissions. At
102
-------
900
800-
700-
600-
Gos Input I960 SCFH
Secondary Air Preheat Temperature As Labeled
15° Spin on Secondary Air
25% Primary Air
Clear Flame
O
E 500-
o.
a.
*
O
2
40O-
-------
a 2$ level of excess O2 (approximately 10$ excess air), the LNO-I burner
emitted 55$ less normalized NO than the IFLB burner. This decrease in
NO emissions dropped to 18$ for NO secondary air preheat.
Flame Luminosity Adjustment
Adjusting the primary air and fuel so that no mechanically induced pre-
mixing occurred prior to ignition (luminous flame) did little to alter the
LNO-I burner emissions without secondary air preheat. Figure 60 illustrates
the normalized NO emissions versus excess O2 for luminous flame operating
conditions. With a 350°C secondary air temperature, the luminous flame had
an NO level, at 2$excess O2, 13$ lower than that of the clear flame. Thus,
the effect of the mixing rate of primary air and fuel on NO emissions is
directly related to secondary air temperature. At ambient combustion air
temperatures, the luminous and clear flames had nearly the same emission
levels, but at the elevated (350°C) combustion air temperature, the clear
flame had the higher emissions.
Variation in Primary Air
To determine the influence of primary air volume on NO formation, the
amount of primary air was reduced to 15$. Figure 61 shows the normalized
NO-versus-excess O2 relationships determined for this primary air volume.
The clear flame with 350°C secondary air preheat displayed little change in
NO emissions between ,15$ and 25$ primary air. The luminous flame was
somewhat more sensitive to primary air volume, yielding a 10$ increase in
the NO emission levels at the lower primary air level. Thus, if a luminous
flame is needed for the desired industrial application, the higher the
percentage of primary air, the lower the NO emissions.
With an ambient secondary air temperature, both flame conditions pro-
duced the lowest levels of NO measured during the burner trial series,
indicating that the percentage of primary air should be reduced as the
secondary air preheat temperature is decreased. To completely generalize
the dependence of NO emissions on primary air volume would require testing
above 25$ primary air; however, because of primary fan capacity, these
trials were not possible.
104
-------
400
300-
6
A
O.
•b
O
•O 200
o
M
6
»_
O
IOO-
Gas Input 1993 SCFH
Secondary Air Temperature
15° Spin on Secondary Air
15 % Primary Air
Clear and Luminous Flames
AS Labeled
365° C
—o
Legend
O Clear Flame
O Cl«qr Flamt
A Luminous Flflme
V Luminous Flame
02 IN
9
FLUE,
Figure 60. Normalized NO concentration as a function of flue O2 for the
LNO-I burner under luminous-flame operating conditions
105
-------
Gas Input 1985 SCFH
Secondary Air Temperature As Labeled
15° Spin on Secondary Air
25 % Primary Air
Luminous Flame
3001
E
a.
ex
o
e
100-
o
-i 1 r—
234
02 IN FLUE. %
T"
5
-i
6
Figure 61. Normalized NO concentration as a function of flue O2 for the
LNO-I burner with 15$ primary air
106
-------
SUMMARY OF BAFFLE BURNER
A synopsis of the operational variables studied and their test results is
presented in Table 6. The technique resulting in the most dramatic reductions
of NO emission levels was external flue-gas recirculation. This verified the
intimate relationship between flarne temperature and NO emissions; thus,
this program's efforts concentrated on reducing flame temperature either by
reducing the rate of combustion or by diluting the flame with combustion
products recirculated within the furnace using combustion aerodynamics.
Although large variations in NO emissions occur because of changes in
the amount of excess air, reduction levels can be established in addition to
their relative effectiveness by comparing the emission levels at several
fixed levels of excess air. The conclusions reached below are based on an
excess air level equivalent to 3 $ oxygen in the flue and a combustion air
preheat temperature of 450°C.
For the intermediate flame length baffle burner (IFLB), which has a
tangential-to-axial velocity component ratio of 0.27:
a. Reducing the firing rate from 3000 SCFH to 2000 SCFH resulted
in a 47 $ reduction in NO emissions.
b. Reducing the combustion air preheat from 450°C to 225°C leads
to a 59$ reduction in NO emissions, and no preheat gives a 78$
reduction.
c. By cooling the walls from 1400°C to 1100°C, the NO emissions
diminished by 55$.
d. External flue-gas recirculation reduced NO emissions by 69$ for
15$ recirculation and 9P$ for 30$ recirculation.
e. Several types of fuel injectors and positions were tested with a
maximum reduction of 67^ measured for the high-momentum
axial nozzle in the throat position. The minimum changes in flame
geometry and luminosity were observed for the divergent nozzle,
which showed a 48$ decrease in NO emissions.
f. The burner block angle was increased from 4-degrees to 8-degrees,
leading to only a 4$ reduction under standard burner operating con-
ditions. However, by also changing the method of fuel injection a
maximum reduction of 60$ was measured with the high-momentum
axial gas nozzle.
107
-------
Table 6. SYNOPSIS OF DATA COLLECTED FOR
THE BAFFLE BURNER
Excess Oxygen, °/0 1 3
Normalized NO, ppm
IFLB Burner (Standard operating conditions: gas input, 3000 SCF/hr;
450 C, secondary air preheat; 2-inch-diameter axial fuel
injector; baffle fuel injector; baffle position; and a 1400 C
wall temperature, 4-degree burner block. )
Standard Operation 390 580
Gas Input, 2000 SCF/hr ^ 250 310
Secondary Air Temperature: 22,5°C 195 235
22°C 110 125
Wall Temperature, 1100 C 225 260
EFGR 15?: 130 180
EFGR 30?'!. 40 60
Radial Nozzle-Throat Position 200 2&0
Axial Nozzle-Throat Position 110 190
Throat Position 280 390
Divergent Nozzle 190 300
Axial Nozzle 210 250
8-Degree Burner Block (B. B. ) 410 560
8-Degree B. B.-Axial Nozzle 200 230
8-Degree B. B.-Divergent Nozzle 220 350
SFLB Burner (Standard operating conditions: gas input, 3000 SCF/hr;
450 C, secondary air temperature; 2-inch-diameter axial
fuel nozzle; baffle position; 8-degree burner block angle;
and 1400 C average wall temperature. )
Standard Operation 450 660
Gas Input, 2000 SCF/hr ' 270 430
Secondary Air Velocity, 125 ft V" 210 250
Secondary Air Temperature: 225°C 165 270
22°C 95 160
Wall Temperature, 1000 C 270 320
EFGR 15ff'o 100 150
EFGR 30^ 60 80
Radial Nozzle 660 800
Divergent Nozzle 440 580
Axial Nozzle 210 390
16-Degree Burner Block (B. B. ) 210 380
16-Degree B. B.-Divergent Nozzle 180 400
16-Degree B. B. -Axial Nozzle 300 390
— Cylindrical Furnace —
IFLB Burner (Standard operating conditions: gas input, 2000 SCF/hr,
350 C, secondary air temperature; 2-inch diameter axial
fuel nozzle, baffle position, 4-degree burner block angle
and 1100 C average wall temperature. )
Standard Operation 460 610
Axial Nozzle 880 1170
Divergent Nozzle 825 1090
LNOX-I 1KO 280
108
-------
For the short flame length baffle burner (SFLB), which has a tangential-to-
axial velocity component ratio of 0.47:
a. Reducing the firing rate from 3000 SCFH to 2000 SCFH resulted
in a 35% decrease in NO emissions.
b. With a 2000 SCFH gas input, the combustion air velocity was in-
creased from 80 ft/s to 125 ft/s, which diminished the NO
emissions by 42%
c. Reducing the combustion air temperature from 450°C to 225°C
resulted in a 59$ NO reduction and, with no preheat, a 76%
reduction.
d. Decreasing the wall temperature from 1400°C to 1000°C reduced
the NO emissions by 52$.
e. Externally recirculating flue gas and blending it with the com-
bustion air produced a 77$ decrease in emissions for 15$
•*• * .I
recirculation and 89% reduction for 30% recirculation.
f. Of the several types of fuel injectors tested, the high-momentum
axial nozzle produced the minimum emissions with a 41%
reduction.
g. The burner block angle was increased from 8-degrees to
16-degrees. This produced a 41% decrease in NO emissions
for all gas nozzles tested.
The IFL/B burner was tested on IGT's cylindrical test furnace, which has
a volume of 226 ft3 and an area ratio between the burner block opening and
the burner wall of 18.2 compared with a 35.9 ratio for the rectangular furnace.
During these trials the following observations were made:
a. At a. 2000 SCFH firing ratio, the cylindrical furnace emissions
were 97% greater than the emissions from the rectangular
furnaces.
b. The high-momentum axial and the divergent nozzles produced
increases in the NO emissions of 92% and 79$, respectively,
although both produced reductions in NO emissions on the
rectangular furnace.
This program has led to the design, development, and testing of a low-NO-
emissions burner (LNOX-I), which, hopefully because of its versatile design,
could be used in a large number of industrial explications. The following
conclusion was drawn from the test data accumulated from LNOX-I:
With 2000 SCFH firing rate on the cylindrical furnace, the LNOX-I
burner resulted in a 54% reduction in NO emissions when compared
with the IFLB burner under similar operating conditions.
109
-------
UTILITY BOILER BURNER
An assembly drawing of the movable-vane boiler burner (MVBB) which
was investigated during this program is shown in Figure 5. The combustion
air enters perpendicular to the axis of the burner and passes through a
register of guide vanes that impart a degree of spin to the air dependent on
the vane orientation. Figure 6 illustrated how the angle of the movable vanes
is measured. The ratio of the average tangential and radial velocity com-
ponents at the exit of the movable -vane register depends only upon the
geometric dimensions of the vanes on the axis perpendicular cross section
(assuming a negligible Reynolds number influence). Leuckel has shown that
this ratio equals —
_ 1 _ tan a /.*
1-B ' 1 +tan atan ( '
where —
B =
cos a
is a blockage factor which accounts for the finite thickness t of the vanes,
and n is the number of vanes in the register. The radial velocity component
at the swirl exit (radius Rx in Figure 6) is given by —
V = _ ^ _ (3)
V oZir^A { '
9
where M is the mass flow rate and A is the axial width of the channels.
Knowing V and CT , the tangential velocity can be determined using the
relationship —
W = crV (4)
There are two fuel injectors commonly used with the MVBB — a "gun"
(a pipe with a hemispherical cap) and a "ring" (two pipes, semicircular in
shape with their ends capped). The gun injector is the least expensive and
therefore was installed with many of the older burners. However, because
of improved flame stability and lower NO emissions, the ring injector is
.X
being installed with most new burners. The total orifice area for an injector
is 0.0491 sq in. for a 1762 SCFH natural gas throughput.
110
-------
The ring injector (Figure 62) has the gas orifices clustered in eight
groups with each group being separated by a 45-degree angle. The number
of holes in a cluster can vary from three to nine. The orifices in a cluster
are of different sizes, the largest hole being located in the center of the
cluster and the smallest at the outside edge of the group. Thus, for the
"No. 3" ring injector there are two different sizes of holes, while the "No. 5"
ring injector has three different sizes of holes. The design criteria for
determining the size of these holes for each injector was supplied by a utility
boiler manufacturer. As an example: for the No. 3 ring injector with a
normal input of 3000 SCFH of natural gas;
• The total orifice area of the ring will be 0.0836 sq in.
\
• The area of the center hole of a.cluster will be (0.072)-(0.0836
sq in.) or 0.0059 sq in. (an approximate diameter of 3/32 inch).
• The area of each of the two smaller holes flanking the center
hole will be (0.0268)'(0.0836 sq in.) or 0.0022 sq in. (an approxi-
mate diameter of 3/64 inch).
* Making the holes with 3/32 and 3/64-inch drill bits would make
a total orifice area of 0.0828 sq in. (less than a 1 $ difference
from the design area of 0.0836 sq in.).
The No. 5 ring injector would have the same total orifice area (0.0836
sq in.) as the No. 3 ring injector because it is also being designed for a
3000 SCFH throughput of natural gas. However, because there are five holes
per cluster instead of three, the cross-sectional area of the holes will be
different. For the No. 5 ring injector —
* The area for the center hole of the cluster will be (0.07)-(0.0836
sq in.) or 0.0058 sq in. (an approximate diameter of 5/64 inch).
• The area of the medium-size holes directly flanking the center
hole will be (0.019)'(0.0836 sq in.) or 0.0016 sq in. (an approxi-
mate diameter of 3/64 inch).
• The two smallest holes, which are located at the outside edges of
the cluster, have an area of (0.0094)-(0.0836 sq in.) or 0.0008
sq in. (an approximate diameter of 1/64 inch).
Thus, for the 3000 SCFH gas input, there is little difference between the
orifices of the No. 3 and the No. 5 ring injectors. Asa result, all the MVBB
tests with a ring fuel injector were conducted using only a No. 3 ring. A
drawing of the injection port arrangement for the No. 3 ring nozzle is -presented
in Figure 62.
Ill
-------
NOZZLE
THE 10° RING NOtZLE. IS SO CALLED BECAUSE
THE C.ENTERLINES OF THS SMALL NOZZLE HOLES
FORM AN AN&LE OF
-------
The gas orifices on a ring injector can be rotated relative to a radial
line passing from the centerline of the burner through the radial axis of the
injector. The normal industrial rotation is 30 degrees toward the combus-
tion chamber. In these trials tests were also conducted at 60-degree and
90-degree rotations.
The gun injector normally has six orifices. These are positioned
symmetrically on a 60 degree conic section relative to the centerline of
the injector. For industrial injectors, there are two different sizes of
orifices, which are arranged in alternating order. The designed area ratio
between the different-sized holes is 0.76, which, for the momentum-scaled
injector to be used in our trials, would require orifice diameters of 8/64
inch and 9/64 inch. Because of the small size difference we decided to use
only the 8/64-inch diameter hole. A drawing of the hemispherical caps used
with the gun nozzle is presented in Figure 63.
A second criterion in fuel injector design is to design the nozzle so that
the fuel will have the desired trajectory when leaving the orifice. For sub-
sonic velocities and incompressible fluids, a path length of 5 to 10 orifice
diameters is required to achieve the proper trajectory. A sonic fluid,
however, will exit with its velocity component at the center of the orifice
normal to the surface tangent. There is a directional velocity distribution
about each orifice. For a sonic fluid, this directional velocity distribution
is independent of wall thickness.
For a utility boiler burner, the fuel injector is designed for critical
sonic flow. Thus, the manufacturer is able to achieve the desired initial
fuel trajectory using thin-walled pipe or tubing at the expense of requiring
a high gas-line pressure. A normal operating line pressure would be 30 psig.
The burner block used during the base-line characterization trials of the
30-degree ring nozzle had a 30-degree divergent angle with a 15.2-inch-
diameter entrance and a 48.2-cm-diameter exit to the furnace. Three nozzle
positions were investigated, as illustrated in Figure 64. When the nozzle is
located 2.5 cm (1 in.) from the burner block exit and is within the 15.2-cm-
diameter refractory duct connecting the burner with the block, the location
is called the throat position. The exit position has the gas injector located
5.1 cm (2 in.) into the divergence of the burner block. When the nozzle is in
the exit position and the deflector plate is positioned 2.5 cm from the tip of
113
-------
CROSS SECTION
OF
NOZZLE HEAD
6 HOLES, f DIAMETER,
EQUALLY SPACED.
MOTEs
DISTANCE FROM CENTERLINE
HEAD TO CENTERLINE, HOLES
AT OUTER SURFACE IS O.8I "
6O° NOZZLE HEAD
CROSS SECTION
OF
NOZZLE HEAD
NOTE:
6 HOLES, i DIAMETER,
EQUALLY SPACED.
DISTANCE FROM CENTERLINE
HEAD TO CENTERLINE HOLES
AT OUTER SURFACE IS O.H7".
3Q° NOZZLE HEAD
Figure 63. Nozzle heads for fuel gun injector
114
-------
I. THROAT POSITION!
£XIT POSITION
3. DEFLECTOR POSITION
DEFLECTOR
Figure 64. Nozale positions tested for
the movable-vane boiler burner
115
-------
the hemispherical cap, the location is called the deflector position. An
illustration of the secondary combustion air deflector plate is given in
Figure 65.
The base-line trials for the 60-degree gun and 30-degree ring nozzles
were conducted with burner operating parameters similar to those found in
industry. These operating conditions included a 470°C secondary combustion
air temperature, a 30-degree vane setting, the exit nozzle position and a
3000 SCFH gas input. Measured CO levels greater than 500 ppm are listed
by concentration next to their corresponding data point. Photographs of
flames produced using these and other operating conditions are presented in
Figures 65a and 65b.
Detailed flame surveys were conducted during the boiler burner trials.
These surveys included flow direction, temperature and gas analysis. The
details of this survey work along with complete in-the-flame data and a
comprehensive listing of the input/output data are presented in Volume II.
However, it is beneficial in understanding the interpretation of the input/
output data to take a brief look at some in-the-flame data.
STANDARD CONDITIONS (60-Degree Gun Nozzle)
For the base-line operating conditions listed above with the 60-degree
gun nozzle; Figure 66 presents the gas species and temperature data
collected along the burner axis. The flow direction data reveals a flow
profile of a. type III. Thus the data in Figure 66 near the furnace front wall
are reflecting properties of the primary recirculation zone. This is born
out by the extremely high temperature measured at the burner block, 1660°C.
The NO concentration is 187 ppm or 55^ of the 340 ppm flue concentration.
Unlike the kiln or baffle burner profiles there is no measurable amount of
methane. The flame length was measured to be 33 cm compared to 186 cm
for the kiln burner and 52 cm for the baffle burner. The NO emission test
results for the 60-degree gun nozzle are plotted in Figure 67. Normalized
NO is given as a function of excess oxygen in the flue and secondary air pre-
heat. Unlike other burner-nozzle combinations at low and intermediate air
preheats (22° and 244°C), the relationship of normalized NO versus flue
oxygen has zero slope.
116
-------
i ia HOLES, £ D\AM.y EQUALLY SPACED.
Figure 65. Secondary combustion air deflector plate
-------
Figure 65a/ 60-degree gun nozzle in exit position
Figure 65b. 30-degree ring nozzle flame in deflector position
118
-------
l<> V A I
TC NO ppm C0%
1700-1 300-1 ' 6
1500-
1300-
250-
200-
IIOO-
900^
700
1
150
100-
50-
DISTANCE FROM BURNER - CM
Figure 66. In-the-flame profiles of boiler
burner operated under typical conditions
-------
60CH
500-
500-
300-
-J
I
200-
100
GAS INPUT 2969 SCFH
NOZZLE EXIT POSITION
30° VANE ROTATION
WALL TEMPERATURE 1340 C
yf BURNER BLOCK ANGLE
AIR PREHEAT AS LABELED
'6000
244°C
123456
Qi IN FLUE, %
Figure 67. Normalized NO concentration as a function of flue O2 for
the movable-vane boiler burner with a 60-degree gun nozzle
120
-------
EXTERNAL FLUE GAS RECIRCULATION (EFGR)
(60-Degree Gun Nozzle)
To determine how successful we can be applying combustion aerodynamics
to recirculate combustion products to the base of the flame as an emission
control technique, we simulated the idealized case (where the combustion air
and products have been thoroughly mixed before ignition) by mixing flue
gases with the combustion air outside the burner. Figure 68 compares the
NO levels as a function of the percentage of oxygen in the flue for 0$, 15$,
and 25$ concentrations of flue-gas recirculation with the 60 -degree gun gas
nozzle. The percentage of flue-gas recirculation (FGR) is determined using
the relationship —
FGR (SCFH) _
A 1UU
(SC
) +
~ Secondary air (SCFH) + Fuel (SCFH)
Again, as was demonstrated for the kiln and ported baffle burners,
external flue-gas recirculation is an extremely effective method of con-
trolling NO emissions from a natural gas flame. The cost of an external
flue-gas system coupled with the additional fan horsepower needed to move
the gas back into the plenum makes its adoption as an industrial emissions
control device undesirable; therefore, it is prudent to use combustion aero-
dynamics to move the flue gas back close to the burner and to mix it as
thoroughly as possible with the air and gas. Referring back to the Flow
Analysis section and Figure 15, the zones containing combustion products,
which for the most part are inert to the kinetics of NO formation, are the
secondary recirculation zone and, to a lesser degree, the primary
recirculation zone.
NOZZLE POSITION (60 -Degree Gun Nozzle)
Our first attempts at altering the combustion aerodynamics to favor
reduced NO emissions for the 60 -degree gun nozzle were to use the three
X.
nozzle positions depicted in Figure 64. These NO emission test results are
illustrated in Figure 69. From flame photographs, we were able to deter-
mine a visual flame length of 76.6 cm. Although the flame length was in-
dependent of the gas nozzle position, the luminosity of the flame was not.
With the gas nozzle in the throat position, the flame was entirely blue (no
luminosity). At the exit position, only the base of the flame was blue, with
121
-------
GAS INPUT 2969 SCFH
NOZZLE EXfT POSITION
30° VANE ROTATION
WALL TEMPERATURE I330°C
30° BURNER BLOCK ANGLE
500
400
I
0s
Kj
-J
1
300-
200-
100-
'6000
1596 FOR
25% FGR
2 3
0. IN FLUE, %
Figure 68. Normalized NO concentration as a function of flue O2 for
the movable-vane boiler burner with a 60-degree gun nozzle,
1. 5% and 2. 5% FGR
122
-------
60 O-i
500-
400-
I
5:
.0
JOO-
200-
100-
GAS INPUT 3004 SCFH
30° VANE ROTATION
WALL TEMPERATURE 1350 c
$0° BURNER BLOCK ANGLE
SECONDARY AIR PREHEAT 460° C
LEGEND
O EXIT POSITION
A THROAT POSITION
V DEFLECTOR POSITION
O, IN FLUE, %
69. Normalized NO concentration as a function of flue O2 for
the movable-vane boiler burner with a 60-degree gun nozzle
in different positions
123
-------
80$ of the flame displaying a high degree of luminosity. The deflector
position displayed a central blue core, with luminosity occurring only at the
flame boundaries. All three nozzle positions should exhibit a type III flow
profile. The deflector plate position is obviously the most desirable as it
shows a decrease in normalized NO of 132 ppm at 2$ excess oxygen or 37 4>
of the concentration measured in the exit position.
NOZZLE TYPE
A*second technique for modifying the combustion aerodynamics is to
vary the velocity components (both direction and magnitude) of the fuel. To
this end, three modified fuel injectors were tested: a low-momentum axial
nozzle (injection velocity of 12 m/s), a high-momentum axial nozzle (in-
jection velocity of 186 m/s), and a divergent nozzle (injection velocity of
199 m/s). The results of these trials are plotted in Figure 70. All of these
data were collected with the nozzles in the throat position. The low-
momentum axial nozzle gave rise to the higher emission levels. This would
imply that the low-momentum nozzle resulted in the highest flame tempera-
tures, with minimum entrainment of recirculating combustion products. The
circumference of the fuel jet was 2.5 cm upon injection, which should indicate
the relative size of the mass-exchange boundary layer between the fuel and
the air. A comparison of the fuel injection velocity (12 m/s, axial) and the
velocity of the combustion air (23.4 m/s, axial; and 18.8 m/s, tangential) in
the throat of the burner block indicates that the mixing is air-momentum
controlled. Thus, the low-momentum nozzle should result in a rapidly mixing
flame, with a very short length and a type III flow profile.
Except at low levels of excess air (below 1.5$ excess oxygen and a
10.5:1 air/fuel ratio), the 60-degree gun and divergent nozzles have similar
levels of NO emissions. Both nozzles inject the fuel with a diverging flow
relative to the burner axis. The gun nozzle introduces the fuel as six jets,
each with a 0.3-cm circumference and a sonic velocity. The divergent nozzle
produces a single jet with a 199 m/s velocity and a 1.2-cm circumference.
Because of the mechanically induced mixing created by a diverging injector
and the resulting void of fuel created on the burner axis, both of these nozzles
will give a type III flow profile.
124
-------
eo;
500
400
I
ox
Q 300
-J
j
200-
100-
O
LEGEND
A LOW MOMENTUM NOZZLE
V DIVERGENT NOZZLE
O 60° GUN NOZZLE
OHIGH MOMENTUM NOZZLE
GAS INPUT 3000 SCFH
NOZZLE THROAT POSITION
30° VANE ROTATION
WALL TEMPERATURE 1360° C
30° BURNER BLOCK ANGLE
SECONDARY AIR PREHEAT 460° C
0. IN FLUE, %
Figure 70. Normalized NO concentration as a function of flue O2 for
the movable-vane boiler burner with a composite
plot of gas nozzles
125
-------
The high-momentum nozzle should produce a type II flame because of
the high axial velocity (186 m/s) of the fuel. The flame resulting from this
nozzle is the longest studied, despite having a larger fuel/air velocity gradient
than the low-momentum axial nozzle. This would indicate that the area of the
mass-exchange boundary layer is the dominant factor in mixing control. This
area is only 1.2 cm for the high-momentum nozzle compared with 2.5 cm for
the low-momentum nozzle. The high-momentum nozzle yields slower mixing,
lower flame temperatures, and lower emissions than the other injectors.
VANE ANGLE (SWIRL) .
(60-Degree Gun Nozzle)
To investigate the effect of secondary combustion air swirl on the com-
bustion aerodynamics and, more importantly, the influence it has on the
formation of NO , trials similar to those discussed previously were con-
Jt
ducted by varying the rotation of the vanes. The orientations studied were
15 degrees, 45 degrees, and 60 degrees. The results of the individual tests
are plotted in Figures 71 through 75. We decided that the most effective way
to compare the results of these tests would be to determine the relationship
between the level of normalized NO emissions and the tangential/radial
velocity ratio. First, a quantitative relationship between the vane orientation,
and the velocity ratio must be made. This is accomplished using Equations 1
and 2 of this section. The results of these calculations are shown in Figure 76.
Table 7 lists the values calculated for a. Note that the measured orientations
of the vanes, listed as «,,, are different than the stated orientation, a. This
is due to an error in fabrication. However, by using Equations 3 and 4, it is
possible to calculate the magnitude of the radial and tangential velocity
components, which are listed as V and W, respectively. Also listed in Table 7
are the normalized NO emissions for the 60-degree gun nozzle in the exit and
deflector positions and the pressure drop in the plenum in inches of water.
(Velocity, NO concentrations, and plenum pressure drops were all quantified
at 2% excess oxygen.)
Figure 77 illustrates the relationship between nor nalized NO and the
tangential/radial velocity ratio for the 60-degree gun in the exit and deflector
positions. The exit position shows a maximum in NO emissions at a velocity
ratio of 1.63 or a 57-degree vane rotation. This gives good agreement with
126
-------
5001
400-
0s
300-
-J
200
100-
GAS INPUT 2994 SCfH
\5° VANE ROTATION
WALL TEMPERATURE
30° BURNER BLOCK ANGLE
SECONDARY AIR PREHEAT
I355°C
457°C
LEGEND
O EXIT POSITION
A THROAT POSITION
V DEFLECTOR POSITION
Q, IN FLUE, %
Figure 71. Normalized NO concentration as a function of flue O2 for
the movable-vane boiler burner with a 60-degree gun nozzle
in different positions and a 15-degree vane angle
127
-------
600-
500-
400-
I
300
-J
i
200-
100-
LEGEND
O EXIT POSITION
A THROAT POSITION
V DEFLECTOR POSITION
GAS INPUT 3976 SCFH
45° VANE ROTATION
WALL TEMPERATURE i34e°c
30° BURNER BLOCK ANGLE
SECONDARY AIR PREHEAT 463°C
02 IN FLUE, °/o
Figure 11. Normalized NO concentration as a function of flue O2 for
the movable-vane boiler burner with a 60-degree gun nozzle
in different positions and a 45-degree vane angle
128
-------
900,
800-
700-
600-
^500-1
0s
^
Cl
400-
I
300-
200-
100
GAS INPUT 3011 SCFH
NOZZLE THROAT POSITION
45° VANE ROTATION
WALL TEMPERATURE I346°C
30° BURNER BLOCK ANGLE
SECONDARY AIR PREHEAT 456°C
LEGEND —
O HIGH MOMENTUM NOZZLE
V DIVERGENT NOZZLE
A LOW MOMENTUM NOZZLE
O 60° GUN N02ZLE
Q, IN FLUE, %
Figure 73. Normalized NO concentration as a function of flue O2 for
the movable-vane boiler burner with a composite plot
of gas nozzles at a 45-degree vane angle
129
-------
500-,
400-
I
0s
3:
8 300
§
~J
I
a:
200-
100-1
GAS \ NPUT 2897 SCFH
60° VANE ROTATION
WALL TEMPERATURE I382°C
30° BURNER BLOCK ANGLE
SECONDARY AIR PREHEAT 46I°C
'12,000
LEGEND
O EXIT POSITION
A THROAT POSITION
V DEFLECTOR POSITION
Q, IN FLUE, %
Figure 74. Normalized NO concentration as a function of flue O2 for
the movable-vane boiler burner with a 60-degree gun nozzle
in different positions and a 60-degree vane angle
130
-------
600i
500'
400-
I
300-
-J
I
200-
100
V,
LEGEND
V DIVERGENT NOZZLE
O GUN NOZZLE
A LOW MOMENTUM NOZZLE
O HIGH MOMENTUM NOZZLE
GAS INPUT ?-883 SCFH
NOZZLE THROAT POSITION
60° VANE ROTATION
WALL TEMPERATURE I376°C
30° BURNER BLOCK ANGLE
SECONDARY AIR PREHEAT 456°C
IN FLUE,
Figure 75. Normalized NO concentration as a function of flue O2 for
the movable-vane boiler burner with a composite plot
of gas nozzles at a 60-degree vane angle
131
-------
9CH
80-
70-
60-
50-
O
K
§ 40-
30-
EO-
10-
I 2 3
TANGENTIAL RADIAL VELOCITY RATIO
Figure 76. Tangential/radial velocity
ratio as a function of vane angle
132
-------
Table 7. LISTING OF BURNER OPERATING CONDITIONS AS A FUNCTION
OF VANE ANGLE FOR THE 60-DEGREE GUN NOZZLE (Gas Input, 2996 SCFH;
Exit and Deflector Positions; 1340°C Wall Temperature; 30-Degree
Burner-Block Angle; 460° C Secondary Air Preheat Temperature)
Normalized NO,
Burner
in H20
15 19 0.37 33.8 12.4 183 197 1.3
30 37 0.81 31.9 25.7 357 225 1.9
45 55 1.55 29.6 46.0 398 247 2.6
60 72 3.54 26.3 93.0 233 227 3.3
a
0.37
0.81
1.55
3.54
V
33.8
31.9
29.6
26.3
W
•Ff / -
It/ S
12.4
25.7
46.0
93.0
Exit
183
357
398
233
pprn
Deflector
197
225
247
227
-------
500-
400-
I
o-
a
300-
100-
O EXIT POSITION
A DEFLECTOR POSITION
GAS INPUT 2996 SCFH
WALL TEMPERATURE IJ40 C
30° BURNER BLOCK AN6LE
SECONDARY AIR PREHEAT 460°C
I 2 3
TANGENTIAL RADIAL VELOCITY RATIO
Figure 77. Normalized NO concentration as a function of tangential/
radial velocity ratio for the movable-vane boiler burner
with a 60-degree gun nozzle
134
-------
the MVBB 53-degree vane rotation determined during the trials previously
conducted under EPA Contract No. 68-02-0216 with air preheats of 22°, 132°,
and 277 °C. From this graph, it is obvious that the vane rotation should be
set as low as possible while maintaining an acceptable flame profile. It is
possible to decrease emissions by going to vane positions above 57 degrees;
however, the plenum pressure drop will make the fan horsepower and
electrical consumption prohibitive.
In-the-flame data for NO emission controlled operating conditions were
collected with the 60-degree gun nozzle in the exit position and a 15-degree
vane angle. Figure 78 gives some of the gas species and temperature data
collected along the burner axis. The flow direction data showed a type II
profile. The primary jet, however, was combustion air and the secondary
jets were fuel. Because of the low tangential velocity the air did not spread
to the edges of the 30-degree burner block but remained on the burner axis.
The 60-degree gun nozzle injected high velocity divergent fuel jets which
penetrated the primary recirculation zones. Despite the primary jet being
mainly combustion air, the temperature at the burner block exit was 1365°C.
The NO concentration at the burner block exit was 20 ppm compared to a
flue concentration of 196 ppm. This flue level of NO emission is a 424
reduction when compared to the emission level from standard operating
conditions with the 60-degree gun.
The gas nozzle position that consistently produced the lowest NO was the
deflector position. The data for this nozzle position are also plotted in
Figure 77. The plot shows little variation in NO as a function of velocity,
with only a 40-ppm difference between minimum and maximum emission
levels. At the 30-degree vane rotation, which is normally used in industry,
there is a 218-ppm normalized emission for the deflector position versus a
357-ppm emission for the exit position. Thus, by using the deflector, it is
possible to reduce the NO by 37$ while maintaining all other furnace and
burner operating parameters unchanged. Photographically, the flame length,
geometry, and luminosity showed little change between these nozzle positions.
However, the photos were taken with a low wall temperature (less than 800°C)
and a negative furnace pressure. Under normal operating conditions, the
flame cannot be defined with either nozzle position.
135
-------
OJ
cr-
5CO-*
4-
O v <
7" /"* A / /O /"* /"; /
' C 'VC/pp/T) C/C/ /o
/ 700-1 300-1 g
/500- 250-
/JOO-1 £00-
I
//OOH
i
I
-j
i
900-1 100-
\
-|
i
700-\ 50-
V
ICC
DISTANCE FROM BURNER - CM
/50
Figure 78. In-the-flame profiles of boiler
burner using 60-degree gun nozzle
-------
Note that the vane rotation producing the maximum NO will be linked to
the cross-sectional exit area and divergent angle of the burner block.
WALL TEMPERATURE (60-Degree Gun Nozzle)
Because of the observed sensitivity of the wall temperature on the flue
concentrations of NO during these tests, we decided to gather data relating
NO to wall temperature during a warm-up cycle. The 60-degree gun nozzle
in the exit position was fired with 3000 SCFH of gas at 3.5^ excess oxygen
and 460°C air preheat. The results are plotted in Figure 79. There is a
dramatic increase in the level of NO emissions above 1300° C. This indicates
that accurate control of the wall temperature is required to produce a con-
sistent set of experimental data. It also confirms that the furnace should be
opened with as low a wall temperature as possible. Although it is not possible
to operate with reduced wall temperature (below 1200°C) in all industrial
applications, these data demonstrates its effectiveness as an NO control
technique.
SECONDARY AIR PREHEAT TEMPERATURE (60-Degree Gun Nozzle)
A similar study was made to determine the sensitivity of NO emissions to
changes in secondary air preheat temperature. This trial was made with a
3000 SCFH gas input, 2<£ excess air, and a 1340°C wall temperature that was
maintained constant throughout the trial by varying the air cooling to the
furnace walls. These test results are plotted in Figure 80. A comparison
of Figures 69 and 80 indicates that reducing the secondary air preheat to
325°C produces the same emission level as using the deflector position with
450°C preheat. Also, by inference from Figure 79, since these data were
collected at 3.5$ excess oxygen, a 1200°C wall temperature with a 450°C air
preheat will produce concentrations of NO similar to those using a 1340°C
wall temperature and a 325°C air preheat.
By use of the data presented in Figures 77, 79, and 80, the level of NO
emissions can even be minimized from a furna.ce that has certain operating
constraints. However, the one operating parameter that seems to override
the amount of secondary air swirl, the furnace wall temperature, and the
secondary air preheat temperature, is the nozzle position. By placing the
injector in the deflector position, the level of e;missions becomes independent
137
-------
60CH
500-
400-
I
Uj 300
200-
100-
GAS INPUT 3007 SCFH
NOZZLE EXIT POSITION
30° VANE ROTATION
30° BURNER BLOCK ANGLE
SECONDARY AIR PREHEAT 455 C
3.5% EXCESS OXYGEN
1000 1100 1200 1300
WALL TEMPERATURE, °C
1400
1500
Figure 79. Normalized NO concentration as a function of wall
temperature for the movable-vane boiler burner with
a 60-degree gun nozzle
138
-------
400'
300-
I
§
i
200-
100-
GAS INPUT 2969 SCFH
NOZZLE EXIT POSITION
30° VANE ROTATION
30 BURNER BLOCK ANGLE
WALL TEMPERATURE 1340 C
2 % EXCESS OXYGEN
-n—O
100
200
300
400
500
600
SECONDARY AIR PREHEAT, °C
Figure 80. Normalized NO concentration as a function of secondary
air temperature for the movable-vane boiler burner
with a 60-degree gun nozzle
139
-------
of swirl. The 60-degree gun nozzle in the exit position can only achieve an
equivalent emission level by a 29$ reduction in secondary air preheat tem-
perature or a decrease in wall temperature of 12%.
STANDARD CONDITIONS (30-Degree Ring Nozzle)
The NO test results for the base-line trials of the 30-degree ring nozzle
X.
are plotted in Figure 81. Data were collected at each of the three nozzle
positions depicted in Figure 64. The throat position had the highest level of
NO emissions. This nozzle position allows for a higher rate of mass exchange
between the fuel and air jets, which results in a higher flame temperature
than the other two nozzle positions. The exit position results in a 40$ re-
duction in the normalized NO at 2$ excess oxygen as compared with the
throat position. This position should minimize the rate of mass exchange
between the fuel and air jets, causing the lowest flame temperature. At 2^
excess oxygen, there is a 54$ reduction in normalized NO as compared with
the throat position and a 24$ reduction as compared with the exit position.
VANE ANGLE (SWIRL)
(30-Degree Ring Nozzle)
Trials were conducted varying the rotation of the vanes to investigate
the effect of secondary combustion air swirl on the combustion aerodynamics
associated with the 30-degree ring nozzle and the influence it has on the
formation of NO . The orientations studied were 15-degrees, 45-degrees,
X.
and 60-degrees. The results of the individual tests are plotted in Figures 82
through 84. Table 8 lists the values calculated for the tangential/radial
velocity ratios in addition to the normalized NO emissions for the 30-degree
ring nozzle in the exit and deflector positions and the pressure drop in the
plenum in inches of water.
Figure 85 illustrates the relationship between normalized NO and the
tangential/radial velocity ratio for the 30-degree ring nozzle in the exit and
deflector positions. The exit position shows a decrease in NO emissions
as the vane rotation is increased. Assuming a linear relationship between
NO emissions, and vane rotation with the nozzle in the exit position, there is
approximately a 1.4 ppm decrease in NO emissions per degree increase in
vane rotation. On the other hand, for every degree increase in vane rotation
there is an additional 0.01-inch H2O pressure drop in the burner plenum.
140
-------
GAS INPUT 2884 SCFH
30° VANE ROTATION
WALL TEMPERATURE )357°C
30° BURNER BURNER ANGLE
400i
I
300-
-J
I
200
100-
LEGEND
O EXIT POSITION
V DEFLECTOR POSITION
A THROAT POSITION
234
0 IN FLUE, %
Figure 81. Normalized NO concentration as a function of flue O2 for
the movable-vane boiler burner with a 30-degree
ring nozzle in different positions
141
-------
400n
300-
i
200
GAS INPUT 2936 SCFH
15° VANE ROTATION
WALL TEMPERATURE
I388°C
30 BURNER BLOCK ANGLE
!00i
1 IOOC
O EXIT POSITION
V DEFLECTOR POSITION
Z 3
0 IN FLUE, %
Figure 82. Normalized NO concentration as a function of flue O2 for
the movable-vane boiler burner with a 30-degree ring nozzle
in different positions and a 15-degree vane angle
142
-------
4 00-,
300-
I
I
200-
100-t
GAS INPUT
2909 SCFH
45° VANE ROTATION
WALL TEMPERATURE 1359° C
30° BURNER BLOCK ANGLE
• LEGEND
O EXIT POSITION
V DEFLECTOR POSITION
L
CL IN FLUE, %
Figure 83. Normalized NO concentration as a function of flue O2 for
the movable-vane boiler burner with a 30-degree ring nozzle
in different positions and a 43-degree vane angle
143
-------
GAS INPUT 2894 SCFH
60° VANE ROTATION
WALL TEMPERATURE
30° BURNER BLOCK ANGLE
1370° C
400-t
I
300-
200-
100-
LECEND
O EXIT POSITION
V DEFLECTOR POSITION
Z 3
0, IN FLUE, %
Figure 84. Normalized NO concentration as a function of flue O2 for
the movable-vane boiler burner with a 30-degree ring nozzle
in different positions and a 60-degree vane angle
144
-------
Table 8. LISTING OF BURNER OPERATING CONDITIONS AS A
FUNCTION OF VANE ANGLE FOR THE 30-DEGREE RING NOZZLE
WITH A 30-DEGREE BURNER BLOCK
(Gas Input, 2907 SCFH; Exit and Deflector Nozzle Positions;
1369°C Wall Temperature; 460°C Secondary Air Preheat; 2$ Excess Oxygen)
Normalized NO,
a
15
30
45
60
"M
19
37
55
72
a
0.37
0.81
1.55
• 3.54
V
ft-/f-
33.8
31.9
29.6
26.3
vv
12.4
25.7
46.0
93.0
Exit
133
122
92
75
Deflector
123
92
126
117
jDunier L!Xr\ ,
in H2O
0.7
0.9
•1.1
1.3
145
-------
400-!
300-
I
zoo-
j
100-
O60° GUN NOZZLE, EXIT POSITION
A 60° GUN NOZZLE, DEFLECTOR POSITION
O30° RING NOZZLE, EXIT POSITION
V30° RING NOZZLE, DEFLECTOR POSITION
TANGENTIAL /RADIAL VELOCITY RATIO
Figure 85. Normalized NO concentration as a function of tangential/
radial velocity ratio for the movable-vane boiler burner
with a composite plot of nozzles
146
-------
The deflector position shows minimum emissions for the 30-degree ring
nozzle at a 30-degree vane rotation. A comparable emission level is not
achieved with the exit position until a 58-degree measured vane rotation is
reached. This would result in an additional 0.28-inch H2O pressure drop.
Included in Figure 85 are data collected under identical operating conditions
except that a 60-degree gun fuel injector was used. The optimum conditions
were the 30-degree ring, the deflector position, and a 30-degree vane rota-
tion. This results in a reduction of 74^ in NO emissions when compared
with the 60-degree gun in the exit position, 58$ when compared with the
60-degree gun in the deflector position, and 25$ when compared with the
30-degree ring in the exit position.
WALL TEMPERATURE (30-Degree Ring Nozzle)
In order to investigate the effect of the temperature on the flue concen-
/
trations of NO, data was gathered relating NO to wall temperature during a
warm-up cycle. The 30-degree ring nozzle in the exit position was fired
with 2856 SCFH of gas at 3 4 excess oxygen and 450° C air prehat. The
results are plotted in Figure 86. There is a dramatic increase in the level
of NO emissions above 1300°C. This indicates that accurate control of wall
temperature is required to produce a consistent set of experimental data.
It also confirms that the furnace should be operated with as low a wall tem-
perature as possible. It is not possible to operate all industrial processes
with reduced wall temperatures (below 1200°C). These data demonstrate the
effectiveness of wall temperature as an NO emission control technique.
SECONDARY AIR PREHEAT TEMPERATURE (30-Degree Ring Nozzle)
A similar study was made to determine the sensitivity of NO emissions to
changes in secondary air preheat temperature. This trial was made with a
2872 SCFH gas input, 3$ excess air, and a 1358°C wall temperature that was
maintained constant by varying the cooling air to the furnace walls. These
trial results are plotted in Figure 87. A comparison of Figures 81 and 87
indicates that reducing the secondary air preheat to 290°C produces the same
emission level as using the deflector position with a 466°C preheat. Also, a
1200°C wall temperature with a 450°C air preheat will produce concentrations
of NO similar to those using a 1358°C wall temperature and a 290°C air
preheat.
147
-------
300n
I
200-
S
-j
1
100-
GAS INPUT 2856 SCFH
NOZZLE EXIT POS'TION
SECONDARY AIR PREHEAT 450° C
3% EXCESS OXYGEN
o
—o-
-O-
-O-
900 1000 1100 1200
WALL TEMPERATURE, °C
1300
KOO
Figure 86. Normalized NO concentration as a function of wall
temperature for the movable-vane boiler burner
with a. 30-degree ring nozzle
148
-------
300-,
200-
I
-J
I
o
100-
GAS INPUT 2872 SCFH
NOZZLE EXIT POSITION
WALL TEMPERATURE
3% EXCESS OXYGEN
I358°C
100
200
300
400
500
SOO
SECONDARY AIR PREHEAT, °C
Figure 87. Normalized NO concentration as a function of secondary
air temperature for the movable-vane boiler
burner with a 30-degree ring nozzle
149
-------
EXTERNAL FLUE GAS RECIRCULATION (30-Degree Ring Nozzle)
A similar study was also made to determine the sensitivity of NO
emissions to changes in external flue-gas recirculation percentage. This
trial was made with a 2942 SCFH gas input, 3$ excess oxygen, a 455°C
secondary air preheat, and a 1365°C wall temperature that was maintained
constant throughout the trial by varying the air cooling to the furnace walls.
These test results are plotted in Figure 88. A comparison of Figures 81
and 88 indicates that having a 9.8^ flue-gas recirculation produces the same
emission level as using the deflector position with a 466°C preheat. Also,
a 1200°C wall temperature with a 450°C air preheat will produce concen-
trations of NO similar to those using a 1315°C wall temperature, 455°C air
preheat, and 9.8$ flue-gas recirculation.
By use of the data presented in Figures 85, 86, 87, and 88, the level of
NO emissions can even be minimized for a furnace that has certain opera-
ting constraints. With the ring nozzle, the level of emissions has become
practically independent of swirl. This level of emissions can be reduced by
lowering the operating wall temperature and/or the level of secondary air
preheat or by externally recirculating flue gases into the secondary combustion
air.
30-DEGREE GUN NOZZLE
A 30-degree gun nozzle was next tested. By decreasing the divergent
angle of the fuel jets from 60 degrees to 30 degrees, it was hoped that this
would delay the fuel/air mixing and allow a larger mass exchange between
the secondary air jet and the secondary recirculation zone. This would result
in a lower peak flame temperature and therefore lower levels of NO emissions.
The experimental data are plotted in Figure 89. Again, as with most previous
input/output data, the deflector nozzle position produced lower levels of NO
emissions than the exit position. The 30-degree gun in the exit position gave
a normalized NO-to-excess oxygen relationship similar to the 60-degree gun
in the deflector position. The minimum emissions for a gun-style nozzle
occurred with the 30-degree gun in the deflector position. At 2% excess
oxygen, it produced a 48$ reduction when compared with the 30-degree gun
in the exit position, a 48$ reduction when compared with the 60-degree gun
150
-------
£
§:
200-
GAS INPUT 2924 S
NOZZLE EXIT POSITION
WALL TEMPERATURE
SECONDARY AIR PREHEAT
3 % EXCESS OXYC-EN
1365° C
455°C
100-t
24
FLUE GAS HECIRCULATION, %
Figure 88. Normalized NO concentration as a function of flue gas
recirculation percentage for the movable-vane boiler burner
with a 30-degree ring nozzle
151
-------
(SOOi
50CH
GAS INPUT 3054 SCFH
30° VANE. ROTATION
WALL TEMPERATURE I390°C
30° BURNER BLOCK ANGLE
400-
I:
5:
pi 300-
200-
100-
V
LEGEND ---
CEXT POSITION
V DE-LECTOR P;
02 IN FLUE, %
Figure 89. Normalized NO concentration as a function of flue O2 for
the movable-vane boiler burner with a 30-degree gun nozzle
152
-------
in the exit position, and a 26$ increase when compared to the 30-degree
ring in the deflector position.
BURNER BLOCK ANGLE
As part of the combustion aerodynamic investigation, a 15-degree and a
45-degree burner block were tested. Figure 90 illustrates the data collected
with a 30-degree vane, a 3044 SCFH gas input, a 1347°C wall temperature, a
460°C secdndary air preheat, and a 15-degree burner-block angle. As with
the 30-degree block, the ring nozzle prodxiced lower NO emission levels
than the 60-degree gun in the deflector position. However, unlike the previous
data, the 60-degree gun in the deflector position produced higher emissions
than the 60-degree gun in the exit position. At 2$ excess oxygen, the 30-
degree ring nozzle produces a 69^ reduction compared with the 60-degree
gun in the exit position and an 81$ reduction when compared with the
60-degree gun in the deflector position.
To investigate the effect of secondary combustion air swirl on the com-
bustion aerodynamics and the influence it has on the formation of NO , trials
ji
were conducted with a 15-degree burner block and varying the rotation of
the vanes. The orientations studied were 15-degrees, 45-degrees, and 60-
degrees. The results of these individual trials are plotted in Figures 91
through 93.
Figure 94 illustrates the relationship between normalized NO and the
.tangential/radial velocity ratio for the 30-degree ring in the exit position,
the 60-degree gun nozzle in the exit position, and the 60-degree gun nozzle
in the deflector position. All three operating conditions showed a minimum
in NO emissions with a 30-degree vane angle. A comparison of Figures 85
and 94 will show the dependence of NO emissions on burner-block angle. The
optimum operating conditions -would be a 15-degree burner block with a
30-degree ring nozzle in the exit position and a 30-degree vane rotation.
The NO emission levels produced by a 45-degree burner block were
tested using a 30-degree ring nozzle. These data are plotted in Figure 95.
Minimum emissions were measured for a 15-degree vane rotation, with
maximum emission being produced by a 45-degree vane rotation.
153
-------
:ocn
500-
400
I
.Q 300
kl
Nl
-J
i
200
100
GAS INPUT 3044 SCFH
30° VANE ROTATION
WALL TEMPERATURE i34/°c
l?° BURNER BLOCK ANGLE
"7500
"27,500
17,000
LEGEND
A60 GUN NOZZLE, EXIT POSITION
V60° GUN NOZZLE, DEFLECTOR POSITION
O30° RING NOZZLE^ EXIT POSITION
02 IN FLUE, %
Figure 90. Normalized NO concentration as a function of flue O2 for
the movable-vane boiler burner with a composite
nozzle plot and a 15-degree burner block angle
154
-------
600i
50CH
400H
sooH
Uj
N
i
a;
i
100
LEGEND
A60°GUN NOZZLE, EXIT POSITION
V60° GUN NOZZLE, DEFLECTOR POSITION
Q300 RING NOZZLE, EXIT POSITION
GAS INPUT 3029 SCFH
15° VANE ROTATION
WALL TEMPERATURE i39o°c
15° BURNER BLOCK ANGLE
2 3
, IN FLUE, %
Figure 91. Normalized NO concentration as a function of flue O2 for
the movable-vane boiler burner with a composite nozzle plot,
a 15-degree burner block angle and a 15-degree vane angle
155
-------
600
500-
400
I
2
Nj
-J
300-
zoo-
100-
o
• LEGEND•
O 30° RING NOZZLE, EXIT POSITION
V 60° GUN NOZZLE, DEFLECTOR POSITION
O500
GAS INPUT 2869 SCFH
45° VANE ROTATION
WALL TEMPERATURE 1395° c
15° BURNER BLOCK ANGLE
02 IN FLUE, %
Figure 92. Normalized NO concentration as a function of flue O2 for
the movable-vane boiler burner with a composite nozzle plot,
a 15-degree burner block angle and a 45-degree vane angle
156
-------
700-
600-
t
a..
"500-
§
I
i
400-
300-
200-
100-
20,000
O
LEGEND
AGO GUN NOZZLE, EXIT POSITION
V 60° GUN NOZZLE, DEFLECTOR POSIT ON
O30° RING NOZZLE, EXIT POSITION
GAS INPUT 2896 SCFH
60° VANE ROTATION
WALL TEMPERATURE
15° BURNER BLOCK ANGLE
I385°C
Q, IN FLUE,
Figure 93. Normalized NO concentration as a. function of flue O2 for
the movable-vane boiler burner with a composite nozzle plot,
a 15-degree burner block angle and a 60-degree vane angle
157
-------
800-
700-
600-
E 500-
8:
cT
^?"
Q 400-
5
§ 300-
200-
IOC-
GXS INPUT Z975 SCFH
WALL TEMPERATURE i390°c
BLOCK ANGLE
O-
LEGEND
O 30° RING NOZZLE, EXIT POSITION
V60° GUN NOZZLE, DEFLECTOR POSITION
A60° GUN NOZZLE/EXIT POSITION
I ? 3
TANGENTIAL!RADIAL VELOCITY RATIO
Figure 94. Normalized NO concentration as a function of tangential/
radial velocity ratio for the movable-vane boiler
burner with a composite nozzle plot
158
-------
500
400-
.
0,300
-o
I
5; 200
100-
O-
—o
LEGEND
A 15° VANE ROTATION
O 30° VANE ROTATION
V45° VANE ROTATION
VANE ROTATION
GAS INPUT 2895 ?CFH
NOZZLE EXIT POSITION
WALL TEMPERATURE i36i°c
45° BURNER BLOCK ANGLE
2 3
0 IN FLUE, %
Figure 95. Normalized NO concentration as a function of flue O2 and
vane angle for the movable-vane boiler burner
with a 30-degree ring nozzle
159
-------
Figure 96 presents plots of normalized NO as a function of burner-block
angle for the four vane rotations investigated. All other vane rotations pro-
duced their minimum emissions with a 30-degree burner block. An extra-
polation of the 30-degree-vane-rotation curve would indicate that the minimum
emissions would be produced with a 0-degree burner-block angle. An attempt
was made at testing the 60-degree gun nozzle with the 45-degree burner
block; however, the flame produced was extremely damaging to the front wall
of the furnace. Flame edges protruded from the observation ports and the
probing doors. Bf.cause of these unacceptable combustion characteristics,
the 60-degree gun nozzle was not tested during the 45-degree burner-block
trial sequence.
Figure 97 presents plots of the normalized NO versus tangential/radial
velocity ratio for the three burner-block angles investigated. The 15-degree
burner block shows a minimum emission level of 67 ppm for a 30-degree vane
rotation. The 30-degree burner block produced a decrease in emissions as a
function of vane rotation for all the angles tested. The minimum level
measured was 75 ppm with a 60-degree vane rotation. The 45-degree burner
block has a peak emission level of 400 ppm at a 45-degree vane rotation. Its
minimum measured level of emissions was 14Z ppm at a 15-degree vane
rotation.
Several gas nozzles were tested with the 45-degree burner block. These
included the 30-degree ring, 30-degree gun, and the low-momentum axial
nozzles. Data for these tests are plotted in Figure 98. All the test data
presented were collected for a 60-degree vane rotation. At excess oxygen
levels below 4.5%, the 30-degree ring and the low-momentum axial nozzles
had comparable levels of emissions. The low-momentum nozzle would be
preferred because it operates at 4-psig line pressure compared with 30 psig
for the 30-degree ring nozzle. The 30-degree gun produced NO levels
approximately 74$ higher than the other two nozzles.
SUMMARY OF BOILER BURNER RESULTS
A synopsis of the operational variables studied and their test results is
presented in Table 9. Although large variations in NO emissions occur
because of changes in the amount of excess air, reduction levels can be
established in addition to their relative effectiveness by comparing the
160
-------
O
GAS INPUT 2950 SCFH
NOZZLE EXIT POSITION
WALL TEMPERATURE 1390° C
SECONDARY AIR PREHEAT 455°
2 % EXCESS OXYGEN
300-
I
-J
200-
100-
LEGEND
A 15° VANE ROTATION
O30° VANE ROTATION
V45° VANE ROTATION
Oeo° VANE ROTATION
—r~
10
—r—
20
—i—
30
—i—
40
—I—
50
—i
60
BURNER BLOCK ANGLE, degrees
Figure 96. Normalized NO concentration as a function of burner
block angle for the movable-vane boiler burner
with a 30-degree ring nozzle
161
-------
400-
300-
§
•-j
I
200-
100-
LEGEND
A 15° BURNER BLOCK
O 30° BURNER BLOCK
O 45° BURNER BLOCK
,o
3AS INPUT 2950 SCFH
NOZZLE EXIT POSITION
WALL TEMPERATURE 1390° c
SECONDARY AIR PREHEAT 4?5
2 % EXCESS OXYGEN
TANGENTIAL. RADIAL VELOCITY RATIO
Figure 97. Normalized NO concentration as a function of tangential/
radial velocity ratio and burner block angle for the movable-vane
boiler burner with a 30-degree ring nozzle
162
-------
600-1
\
100-
LEGEND
A 30° GUN NOZZLE
V LOW MOMENTUM AXIAL NOZZLE
O 30° RING NO/ZLE
GAS INPUT 3041 SCFH
NOZZLE EXIT POSITON
60° VANE ROTATION
WALL TEMPERATURE i396°c
45° BURNER BLOCK ANSLE
On
IN FLUE, °/o
Figure 98. Normalized NO concentration as a function of flue O2 for
the movable-vane boiler burner with a composite nozzle
plot, 60-degree vane angle and 45-degree burner block
163
-------
T able 9. SYNOPSIS OF DATA COLLECTED FOR
THE BOILER BURNER
Excess Oxygen, % 1 3
• Normalized NO, ppm
(Standard operating conditions: gas input 2960 SCFH; secondary air preheat
450°C; 60-degree gun nozzle; vane angle 30-degrees; exit nozzle position;
wall temperature 1340°C and 30-degree burner block angle)
Standard Operation 310 400
Secondary Air Temperature
244°C 160 163
22°C 63 93
EFGR, 15$ ' 100 137
EFGR, 25^ 27 33
Throat Position 343 423
Deflector Position 173 277
Throat Position
Low Axial Momentum Nozzle 370 512
Divergent Nozzle 280 450
Axial Nozzle 210 315
15-Degree Vane Angle
Exit Position 173 197
Throat Position 187 188
Deflector Position 180 213
45-Degree Vane Angle
Exit Position 328 465
Throat Position 267 283
Deflector Position 223 300
Throat Position, 45-Degree Vane Angle
Low Axial Momentum Nozzle 335 470
Divergent Nozzle 385 580
Axial Nozzle 475 685
60-Degree Vane Angle
Exit Position 220 270
Throat Position 253 310
Deflector Position 187 280
Throat Position, 60-Degree Vane Angle
Low Axial Momentum Nozzle 250 343
Divergent Nozzle 437 555
Axial Nozzle 163 237
30-Degree Gun Nozzle
Exit Position 175 293
Deflector Position 90 183
15-Degree Burner Block, Deflector Position
15-Degree Vane Angle 520 610
30-Degree Vane Angle 327 393
45-Degree Vane Angle 317 477
60-Degree Vane Angle 625 730
164
-------
Table 9, Cont. SYNOPSIS OF DATA COLLECTED FOR
THE BOILER BURNER
(Standard operating conditions: gas input 2938 SCFH; secondary air preheat
450°C; 30-degree ring nozzle; exit nozzle position; wall temperature 1380°C;
30-degree vane angle and 30-degree burner block)
Standard Operation 93 173
Throat Position 137 247
Deflector Position 53 128
15-Degree Vane Angle
Exit Position 118 153
Deflector Position 90 164
45-Degree Vane Angle
Exit Position 68 107
Deflector Position 100 153
60-Degree Vane Angle
Exit Position 58 102
Deflector Position 85 160
EFGR, 12$ -- 108
EFGR, 24$ -- 28
15-Degree Burner Block Angle
15-Degree Vane Angle 100 210
30-Degree Vane Angle 45 92
45-Degree Vane Angle 150 293
i 60-Degree Vane Angle 325 537
45-Degree Burner Block Angle
15-Degree Vane Angle 129 171
30-Degree Vane Angle 154 200
45-Degree Vane Angle 311 456
60-Degree Vane Angle 209 311
165
-------
emission levels at several fixed levels of excess air. The conclusions
reached below are based on an excess air equivalent to 3 $ oxygen in the flue
and a combustion preheat temperature of 450°C.
For the 60-degree gun nozzle and a 30-degree vane angle, the following
conclusions can be drawn:
a. Reducing the combustion air preheat from 450°C to 244°C leads
to a 59$ reduction in NO emissions, and no preheat gives a 77$
reduction.
b. External flue-gas recirculation reduced NO emissions by a 66$
jf jt ji
for 15% recirculation and 91* for 25% recirculation.
c. Several nozzle positions were tested with a maximum reduction
measured for the deflector position of 31$. The nozzle type
which produced the lowest level of emissions for the throat
position was the axial nozzle with a 21$ reduction.
d. NO emissions showed a maximum for a 57-degree vane rotation.
The test results show that the vane rotation should be set as low
as possible while maintaining an acceptable flame profile.
For the 30-degree ring nozzle and a 30-degree vane angle, the following
conclusions can be drawn:
a. At standard operating conditions, the 30-degree ring produced
57$ fewer NO molecules than the 60-degree gun.
b. The deflector nozzle position results in a 26$ reduction in NO
emissions.
c. External flue-gas recirculation reduced NO emissions by 38$
for 12$ recirculation and 84$ for 24$ recirculation.
d. The combination of burner block angle and vane rotation which
resulted in minimum emissions were: a 15-degree burner
block and a 30-degree vane rotation producing a 56^ reduction
in NO emissions; or a 30-degree burner block and a 60-degree
vane rotation which resulted in a 48$ reduction.
In conclusion, the 30-ring nozzle consistently produced lower levels of
NO emissions than the 60-degree gun nozzle. Additional reductions are
possible by adapting the deflector or nozzle position or by using a 15-degree
burner block and a 30-degree vane angle. The emissions from the 15-degree
burner block and the 30-degree vane angle are even lower than those
measured using 12$ EFGR.
166
-------
APPENDIX. Data Correlation
SUMMARY
This project generated a large quantity of data on the effect of
numerous operating variables on pollution characteristics of three differ-
ent burner types. Many conclusions can be drawn on the basis of the
data alone. It is useful, however, to attempt to derive some more
general conclusions on the effects of burner parameters on pollutant
emissions. This can best be accomplished, at this time, by analyzing
and correlating some of the data, in the light of specific theories or
questions. In this work the focus was on correlating the influence of
flue gas recirculation and air preheat on NO emissions as a function of
the adiabatic flame temperature. Additionally, attempts were made to
correlate the effects of excess air levels.
Results indicated that the data, could be correlated to a remarkable
extent. Indeed, it appears that (provided the restriction of unchanged
fluid dynamics is maintained) the effects of three variables — air preheat,
flue gas recirculation and excess air — for a given burner, can be
(roughly) predicted from a single data point. These results are intriguing,
and should provide impetus for further interpretation in the light of
fundamental phenomena. The practical implications are important too,
since our results indicate that given experimental emission measurements
on a particular natural gas-fired system at different air preheat temper-
atures, it is possible to conservatively estimate the amount of flue gas
recirculation needed to meet specific NO emission requirements. This,
in turn, allows a reasonable cost/effectiveness analysis prior to system
procurement and installation.
INTRODUCTION
Pollutant emissions from the diverse furnace flames investigated in
this project arise through the interaction of many complex physical and
chemical phenomena. Ideally, one would like to interpret the experi-
mental data obtained in the light of these physical or chemical phenomena.
However, a rigorous theoretical analysis is difficult and outside the scope
of this project. It was felt, therefore, that interpretation of the results
could best be achieved, through attempts to correlate the data in a
rational but empirical way.
167
-------
Because of the complexity of the process involved, and the volume
of the data acquired, it is first necessary that the scope of this analysis
portion of the project be narrowed. The focus, therefore, is on a single,
limited aspect of the problem; namely, the relationship between air pre-
heat and flue gas recirculation (FGR). The objective is to show that for
a given piece of burner hardware, secondary air preheat and flue gas
recirculation both affect NO emissions, primarily by changing the local
.X
flame temperature which is assumed to be directly related to (but probably
not equal to) the theoretical adiabatic temperature of the flame.
Inherent in this analysis is the assumption that the temperature of
the reaction zone in which NO is being formed in the flame, can have
some "average" value which is directly proportional to the adiabatic
flame temperature.
APPROACH
In this section the general approach used throughout the analysis is
described. Discussion of specific experimental results is deferred to
the next section. The following burner types were examined:
• Kiln burner — combination nozzle
• Intermediate flame length ported baffle burner — standard nozzle
• Short flame length ported baffle burner — standard nozzle
• Movable-vane boiler burner — gun nozzle.
Data sets included various secondary air preheats and flue gas
recirculation levels for each burner. All the data in a particular set,
however, were for the same fuel injector types and position, quarl angle,
excess air, firing rate, etc. so that the fluid dynamics would be approx-
imately the same.
Three types of plots were made for each burner:
1.0. NO vs.
The NO emissions (in ppm, dry, normalized to stoichiometric) were
plotted against the adiabatic flame temperature based on the particular
air preheat, FGR level and excess air of the test point. In this manner,
data at various preheats and FGR levels could be superimposed on the
168
-------
same graph, parametric with excess air. If flue gas recirculation and
decreased air preheat both reduce NO via a purely thermal mechanism,
j£
and if the local flame temperatures are directly related to the adiabatic
flame temperature, then all the data for a particular burner at a given
excess air should fall on a smooth curve.
2.0. Ln(NO) vs. 1 /
Secondly, the NO emissions (in ppm, dry, normalized) were plotted
vs. the reciprocal adiabatic flame temperature on a semi-log basis in
an attempt to obtain simple semilogarithmic correlations. Such a cor-
relation would indicate that the rate determining mechanism has an
Arrhenius dependence on temperature. Originally the theoretical flame
temperature was calculated allowing for the high temperature disassoci-
ation of both CO2 and water vapor according to the following equilibrium
relations:
1/2 O?. + CO = CO2
1/2 O2 + H2 = H2O
The correlation was improved; however, when this partial equilibrium
was not taken into account; i.e., when the fuel was assumed to burn
completely to CO2 and H2O. In retrospect, perhaps this is not surprising.
Clearly the actual, local flame temperatures are below the adiabatic
temperature and at these lower temperatures the equilibrium would heavily
favor CO2 and H2O. For this reason, all of the adiabatic temperatures
shown in the following sections assume complete combustion to CO2 and
H20.
3.0. Ln(NO/[02]) vs. 1 /
Finally, the NO emissions (wet, as measured) divided by the percent
oxygen (wet) in the flue gas were plotted vs. the reciprocal flame tem-
perature on a semi-log basis.
If one assumes that NO formation is via the Zeldovich mechanism —
N2 + O = NO + N
O2 + N = NO + O
169
-------
and that oxygen atoms are equilibrated —
M + O2 = O + O + M
then it can be shown that —
d[NO]/dt a exp[-A/T] * [Ozl1/2
If one further makes the gross (unsupported) assumption that the resi-
dence time/temperature history allows direct integration, then a plot of—
ln{NO/[02]1/2} vs. 1/T
should be linear. The data considered herein do not support this;
however, they correlated remarkably well when plotted as —
ln{NO/[02]} vs. 1/
This suggests that a simple equilibrium relationship between oxygen
atoms and molecules is invalid for the reaction zone in which NO is
being formed. It should be noted that for diffusion flames, at any
instant, the midpoint of the reaction zone is determined by that point
at which the overall stoichiometric ratio is equal to one. Furthermore,
the linear dependence on excess oxygen is intriguing, and suggests that
the thickness of the reaction zone (or apparent volume in which NO is
formed) is controlled by diffusion of oxygen. Further work is necessary
to fully interpret these results.
DETAILED RESULTS
Table 10 gives the specific details of each of the burners considered.
These cases were selected because there was both air preheat and flue
gas recirculation data available. Three excess oxygen levels were used
throughout the analysis: 1.5, 3.0, and 4.5% O2. Where necessary, data
were obtained at exactly these levels by interpolating between actual
experimental points.
In all of the plots presented in this section, the parametric lines
represent the different levels of excess oxygen. The circled points are
data with no flue gas recirculation while the diamonds represents points
with flue gas recirculation.
170
-------
Table 10. BURNER OPERATING CONDITIONS
USED IN DATA CORRELATION
BURNER NO. DESCRIPTION
Kiln burner - Combination nozzle
810 SCFH axial gas
1890 SCFH radial gas
3.5% primary air
1130°C wall temperature
22°C primary air temperature
( Table 43, Volume II)
Intermediate flame length
ported baffle burner - Standard
gas nozzle
3070 SCFH gas
baffle gas nozzle position
1435°C wall temperature
4° burner block angle
( Table 57, Volume II)
Short flame; ported baffle burner -
Standard gas nozzle
3070 SCFH gas
baffle nozzle position
1360°C wall temperature
8° burner block angle
( Table 74, Volume II)
Movable-vane boiler burner
60° gun nozzle
2969 SCFH gas
exit gas nozzle position
30° vane angle
1340°C wall temperature
30° burner block angle
( Table 82, Volume II)
171
-------
Figures 99 to 101 show the results for the kiln burner. Figure 99 indi-
cates that the FGR data tend to fall slightly above the normal curve. This
is not typical of the other burners and may be related to the fact that the
kiln burner has a primary air stream which is not supposedly preheated.
One possible explanation for FGR being "less effective" than reduced air
preheat is that since the FGR data are at a high preheat, there may have
been some "artificial heating" of the primary air stream prior to combustion.
Figure 100 again shows the FGR data to be slightly higher. Figure 101 shows
all the data for the kiln burner plotted as NO/[O2] vs. I/TADB and considering
the data are for different levels of —
• Excess air
• Flue gas recirculation
• Secondary air preheat
the linearity is quite remarkable.
Figures 102 through 104 show the results for the intermediate flame
length ported baffle burner. As Figure 102 shows, the preheat and FGR
data generally fall on the same curve for a particular excess air. Figure 104
again shows good linearity with respect to first order oxygen dependence and
reciprocal adiabatic temperature.
Figures 105 through 107 show the results for the short flame length
ported baffle burner. As Figure 105 indicates, the FGR data for this burner
fall far below the 0% FGR line. This type of burner produces a very short,
rapidly mixed, intense flame. Perhaps in this case the FGR is being more
beneficial than expected because in addition to reducing local flame tempera-
tures directly, it is spreading the entire flame out. It should be recalled
that the secondary air injection velocity is different by almost a factor of two
between the circle and diamond points at the same flame temperature. For
example, the point labeled A in Figure 105 is at 210°C preheat, 0% FGR;
while the point labeled A1 is at 455°C preheat and 15% FGR. Thus, it is not
surprising that a difference in fluid dynamics might exist. This is further
substantiated by the original plot of the data ( Figure 51) where the high pre-
heat data exhibit a different dependence on excess air than the no preheat
data). Figure 107 shows the composite plot and the agreement is not
particularly good. Again, the FGR data are lower than would be predicted.
172
-------
70O-r
600-
1
cf
i
N
1
500-
400-
200-
100
LEGEHD
v 13%
o 0%
1700
1300
2100
2300
2500
ADIABATIC TEMPERATURE °K
Figure 99. Normalized NO concentration as a function of
adiabatic flame temperature for the kiln burner
-------
1000-
-J
I
LKE ND —
v /3% £76 fl
0 0%
0.43
0.47
0.51
\
0.55
Figure 100. Normalized NO concentration as a function of
[1000/T.r.ia( °K) ] for the kiln burner
-A.JLJ 13
174
-------
/OO-i
8-
IO-\
LEGEND
V 13%
0 0% EFGR
43
.47
.51
55
Figure 101. Logarithmic ratio of NO/[O2] as a function of
[1000/TADB( °K) 3 for the kiln burner
175
-------
Q.
or
i
Ci
700-i
6OO-
5OO-
400-
3OO-
200-
IOO-
v 13% EFGR
0 0%
1700
T
1900 2/00 Z3OO
ADfABATIC TEMPERATURE °K
2500
Figure 102. Normalized NO concentration as a function of adiabatic flame temperature
for the intermediate flame length ported baffle burner
-------
IQOO-i
I
CL
IOO-
-J
I
s
/O-
LEGEHD
v
0
45
I
.47
i
.57
.55
/000/TA
'AOB
Figure 103.
[1000/T
Normalized NO concentration as a function of
( °K) ] for the intermediate flame length
ported baffle burner
177
-------
1000
v 13% EF6R
O 0% EFBR
.43
/000/rnfl ftc")
.55
Figure 104. Logarithmic ratio of NO/[O2] as a function of
[1000/T ( °K) ] for the intermediate flame length
ported baffle burner
178
-------
SO
7OO-i
600-
500-
s- 400-
<:
a
1
300-
200-
IOO-
LEGEHD-
v 13% £FGR
00% EFGR
1700
I90O
2100
2300
2500
ADfABATtC TEMPERATURE °K
Figure 105. Normalized NO concentration as a. function of
adiabatic flame temperature for the short flame ported baffle burner
-------
/OOO-i
Q_
Q.
S"
IOO-
o:
LEGEHD-
v
0 0%
.43
.47
.51
.55
fOOO/TA
'ADB
Figure 106. Normalized NO concentration as a function of
[1000/TArm( °K) ] for the short flame ported baffle burner
180
-------
/OOO-i
v
o o%
.55
Figure 107. Logarithmic ratio of NO/[O2] as a function of
[1000/TADB( °K) ] for the short flame ported baffle burner
181
-------
Figures 108 through 110 show the results for the movable-vane boiler
burner. As Figures 108 and 109 indicate, the low FGR level ( 15%) falls
very nicely on the solid curves. The higher FGR data (25%) are below the
curves, again possibly due to some type of fluid dynamic effects. The
composite plot, Figure 110, shows reasonable agreement except at the
very high levels of flue gas recirculation.
DISCUSSION
As the previous section has shown, in many cases, emission data for
various levels of air preheat and FGR fall on a smooth curve when plotted
against adiabatic flame temperature. In no case did the FGR emissions
exceed the smooth curve through the no-preheat data by more than 20 ppm.
Thus, given experimental emission measurements on a particular natural-
gas-fired system at different air preheat temperatures, it is possible to
conservatively estimate the effect of installing a certain size FGR system.
This allows a reasonable cost/effectiveness analysis prior to system
installation.
By far the most unusual and unexpected result of the analysis is the
finding that the vast majority of the data can be linearized in the form
ln{NO/[O2]} vs. !/TADB. Further, while each plot has a different inter-
cept, they have identically the same slope for all of the burners examined.
This is particularly surprising since the kiln burner gives a fifteen-foot-
long, very thin flame and relatively low NO ( always below 300 ppm) while
the short flame ported baffle burner gives a flame less than 2 feet long
and emissions over 700 ppm in some cases.
CONCLUSIONS
• The effect of flue gas recirculation on NOX emissions can generally
be correlated (or estimated) to within 20% for a wide variety of
burners if one has data at various preheat levels and no FGR.
• Adiabatic flame temperature is a reasonable parameter to correlate
preheat and FGR data against.
• Further study should be given to the significance of the high degree
of correlation obtained between ln( NO/O2) vs. 1/T.,-.,,.
182
-------
oo
_
0.
700-,
600-
500-
400-
§
1"°
.<£
< 200
/00-
v 13% EF6R
O 0% EfGR
1700
I90O
2100
2300
2500
AD/ABATIC TEMPERATURE °K
Figure 108. Normalized NO concentration as a function of
adiabatic flame temperature for the movable-vane boiler burner
-------
/OOO-i
a.
s
IOO-
I
s
\
V 13% EF6R
o 0%
10-
T
.57
.47
t°00/TAD8
.55
Figure 109. Normalized NO concentration as a function of
[1000/TADB( °K) ] for the movable-vane boiler burner
184
-------
/00-
- n
\
O
00% EFGR
.43
.47
i
.51
/000/TA
'ADB
.55
Figure 110. Logarithmic ratio of NO/COz] as a function of
[1000/T _ ( °K) ] for the movable-vane boiler burner
AJ-/Jj
185
-------
CONVERSION TABLE
ENGLISH TO SI METRIC CONVERSION FACTORS
To Convert
Lb/106 Btu
106 Btu/hr
PSI
SCFH
Ft/s
Inch
Feet
Feet2
From
Inches of water (pressure)
Lb/ft3
GPM
Inch2
°F
J ~~
g
s =
MWt
GPM
Pa
m =
k
n =
M
C
F =
PSI
SCFH
Joule
gram
second
Megawatts thermal
To
ng/J
MWt
Pa
m3/s
m/s
m
m
m2
Pa
kg/m3
m3/s
m2
°C
Multiply By
4. 299 E + 02
2. 928751 E -01
6. 894757 E +03
7. 865790 E -06
3. 048000 E -01
2. 540000 E -02
3. 048000 E -01
9. 290304 E -02
2. 4884 E + 01
1. 601846 E +01
6. 309020 E - 05
6.451600 E -04
t°c=(t°F- 32) /I. 8
Gallons ( U. S. liquid) /minute
Pascal
metre
kilo ( 103)
_ 9
nano (10 )
mega ( 106)
Celsius
Fahrenheit
pounds per square inch
standard cubic feet per
hour
186
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-76-098a
2.
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
Burner Design Criteria for Control of NOx from
Natural Gas Combustion; Volume I. Data Analysis
and Summary of Conclusions
6. REPORT DATE
April 1976
6. PERFORMING ORGANIZATION CODE
. AUTHOR(S)
D.R. Shoffstall
8. PERFORMING ORGANIZATION REPORT NO
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Institute of Gas Technology
IIT Center, 3424 South State Street
Chicago, Illinois 60616
10. PROGRAM ELEMENT NO.
1AB014; 21BCC-029
11. CONTRACT/GRANT NO.
68-02-1360
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT ANl
Final; 6/73-9/75
NO PERIOD COVERED
14. SPONSORING AGENCY CODE
EPA-ORD
is. SUPPLEMENTARY NOTES Project officer for this report is D.G. Lachapelle, Mail Drop 65.
Ext 2236.
is.ABSTRACT Volume I of the report gives details of, and analyzes, trials conducted with
natural gas to determine the relationship between combustion aerodynamics and pol-
lution emission characteristics of industrial burners. Three burner types were stu-
died (kiln, ported baffle, and movable vane boiler), based on relative gas load and
estimated total industrial emissions. Experimental measurements on a pilot-scale
furnace included baseline characterization of each burner and variation of primary
operating parameters (air preheat, air/fuel ratio, firing rate, heat release rate,
position of gas nozzle in burner block, and air swirl intensity). Additional emissions
data were gathered for suspected control conditions (fuel injector design, flue gas
recirculation, fuel/air momentum ratio, and burner block angle). It also describes
in detail the experimental facility and sampling probes used to collect the data.
Volume n discusses completely the procedurexused to select the test burners. It
includes detailed flame characterizations of baseline operations assembled from in-
:he-flame temperature, gas species, and flow direction data analysis. Similar in-the-
lame studies were made for control conditions which minimized emissions for each
burner type. It also includes all raw data collected from the input/output trials.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Air Pollution
Nitrogen Oxides
Aerodynamics
Natural Gas
ombustion Control
Burners
Flames
Swirling
Air Pollution Control
Stationary Sources
Axial Injection
Radial Injection
Swirl
Industrial Burners
13B
07B
20D
2 ID
21B
13A
13H,07A
8. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
212
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
187
------- |